Search

Today the reigning opinion about computer proof assistants based on constructive logic (even from some of the developers of these tools!) is that, while they are very helpful for doing math, they are an absurdly heavy-weight solution to use for practical programming. Yet the Curry-Howard isomorphism foundation of proof assistants like Coq [Yves Bertot and Pierre Castéran. Interactive Theorem Proving and Program Development. Coq'Art: The Calculus of Inductive Constructions. Texts in Theoretical Computer Science. Springer Verlag, 2004] gives them clear interpretations as programming environments.My purpose in this position paper is to make the general claim that Coq is already quite useful today for non-trivial certified programming tasks, as well as to highlight some reasons why you might want to consider using it as a base for your next project in dependently-typed programming.

We describe the experience of implementing a Domain-Specific Language using transformation to a General Purpose Language. The domain of application is image processing and low-level computer vision. The transformation is accomplished using the Stratego/XT language transformation toolset. The implementation presented here is contrasted with the original implementation carried out many years ago using standard compiler implementation tools of the day. We highlight some of the unexpected advantages afforded to us, as language designers and implementers, by the source-to-source transformation technique. We also present some of the practical challenges faced in the implementation and show how these issues were addressed.

As suggested by the title of Shoham, Powers, and Grenager's position paper [Y. Shoham, R. Powers, T. Grenager, If multi-agent learning is the answer, what is the question? Artificial Intelligence 171 (7) (2007) 365–377, this issue], the ultimate lens through which the multi-agent learning framework should be assessed is “what is the question?”. In this paper, we address this question by presenting challenges motivated by engineering applications and discussing the potential appeal of multi-agent learning to meet these challenges. Moreover, we highlight various differences in the underlying assumptions and issues of concern that generally distinguish engineering applications from models that are typically considered in the economic game theory literature.

Nowadays, with the rapid development and wide application of Internet of Things technology, Industrial Internet of Things (IIoT) has become an essential part of industrial systems. A large number of IoT devices has enabled the diversification and decentralization of energy systems, and the Energy Internet makes it possible to interconnect distributed energy sources and consumers. However, as the number of entities involved in the energy transactions grows and the forms of transactions diversify, the traditional centralized energy trading model faces the problems such as insufficient scalability, vulnerabilities to cyber-attack, and low processing efficiency. To address these challenges, we propose BC-ETS, a secure and efficient BlockChain based Energy Trading Scheme. In BC-ETS, the energy trading model is divided into two levels, which can not only protect privacy but also achieve the balance between power supply and demand. Additionally, to adapt to the relatively weak computing power of the underlying IoT equipment in energy internet, a credibility-based equity proof mechanism is designed to greatly improve the system availability. The analysis shows that BC-ETS can meet the security requirements and has a better performance compared with other similar energy trading schemes. - 2019 Elsevier B.V. The work is partially supported by the National Natural Science Foundation of China under Grant 61972148 , Beijing Natural Science Foundation under grant 4182060 . Zhitao Guan (M'13) is currently an Associate Professor in the School of Control and Computer Engineering, North China Electric Power University. He received his BEng degree and PhD in Computer Application from Beijing Institute of Technology, China, in 2002 and 2008, respectively. His current research focuses on smart grid security, wireless security and cloud security. Dr. Guan has authored over 30 peer-reviewed journal and conference papers in these areas. He is a Member of the IEEE. Xin Lu is currently a Master candidate in the School of Control and Computer Engineering, North China Electric Power University (NCEPU). He received his BEng degree from NCEPU in 2017. His current research focuses on smart grid security and cloud security. Naiyu Wang is currently an undergraduate student at the School of Control and Computer Engineering, North China Electric Power University. Her current research focuses on smart grid security and cloud security. Jun Wu (S'08-M'11) is an Associate Professor of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, China. He obtained his PH.D. Degree in Information and Telecommunication Studies (GITS) at Waseda University, Japan. He was a postdoctoral researcher for the Research Institute for Secure Systems (RISEC), National Institute of Advanced Industrial Science and Technology (AIST), Japan, from 2011 to 2012. He worked as a researcher for the Global Information and Telecommunication Institute (GITI), Waseda University, Japan, from 2011 to 2013. His research interests include the advanced computation and communications techniques of smart sensors, wireless communication systems, industrial control systems, wireless sensor networks, smart grids, and more. He has hosted and participated in several research projects for the National Natural Science Foundation of China, National 863 Plan and 973 Plan projects, etc. He has been a Guest Editor for the IEEE Sensors Journal and a TPC Member of several international conferences including WINCON 2011, GLOBECOM 2015, etc. He is a member of IEEE. Xiaojiang Du (M'04-SM'09) is currently a professor in the Department of Computer and Information Sciences at Temple University. Dr. Du received his B.S. and M.S. degree in electrical engineering from Tsinghua University, Beijing, China in 1996 and 1998, respectively. He received his M.S. and Ph.D. degree in electrical engineering from the University of Maryland College Park in 2002 and 2003, respectively. Dr. Du was an Assistant Professor in the Department of Computer Science at North Dakota State University between August 2004 and July 2009, where he received the Excellence in Research Award in May 2009. His research interests are security, wireless networks, computer networks and systems. He has published over 200 journal and conference papers in these areas. Dr. Du is a Senior Member of IEEE and a Life Member of ACM. Mohsen Guizani (S'85�M'89�SM'99�F'09) received the B.S. (with distinction) and M.S. degrees in electrical engineering, the M.S. and Ph.D. degrees in computer engineering from Syracuse University, Syracuse, NY, USA, in 1984, 1986, 1987, and 1990, respectively. He is currently a Professor at the Computer Science and Engineering Department in Qatar University, Qatar. Previously, he served in different academic and administrative positions at the University of Idaho, Western Michigan University, University of West Florida, University of Missouri-Kansas City, University of Colorado-Boulder, and Syracuse University. His research interests include wireless communications and mobile computing, computer networks, mobile cloud computing, security, and smart grid. He is currently the Editor-in-Chief of the IEEE Network Magazine, serves on the editorial boards of several international technical journals and the Founder and Editor-in-Chief of Wireless Communications and Mobile Computing journal (Wiley). He is the author of nine books and more than 500 publications in refereed journals and conferences. He guest edited a number of special issues in IEEE journals and magazines. He also served as a member, Chair, and General Chair of a number of international conferences. Throughout his career, he received three teaching awards and four research awards. He also received the 2017 IEEE Communications Society WTC Recognition Award as well as the 2018 AdHoc Technical Committee Recognition Award for his contribution to outstanding research in wireless communications and Ad-Hoc Sensor networks. He was the Chair of the IEEE Communications Society Wireless Technical Committee and the Chair of the TAOS Technical Committee. He served as the IEEE Computer Society Distinguished Speaker and is currently the IEEE ComSoc Distinguished Lecturer. He is a Fellow of IEEE and a Senior Member of ACM. Scopus

Modern theory of safety deals with systemic approach to safety, formalized in form of several systemic prediction models or methods such as FRAM (Functional Resonance Analysis Method) or STAMP (System-Theoretic Accident Model and Processes). The theory of each approach emphasizes different viewpoints to be considered in approaching various industrial safety issues. This paper focuses on FRAM and its functional viewpoint for modern complex sociotechnical systems. The methodology in this paper is based on the utilization of foundational ontologies to conceptualize the core ideas of FRAM, with the focus on the concept of functions as used in theory. The outcomes of the case study in the aviation domain provide for what needs to be determined to properly model functions in FRAM and they allow for better utilization of the method in real-case applications. The results also confirm some previous research, suggesting that modern systemic approach to safety is theoretically grounded on common - or at least complementary - tenets, to be prospectively integrated by means of ontology engineering.

COVID-19 is a buzz word nowadays. The deadly virus that started in China has spread worldwide. The fundamental principle is “if the disease can travel faster information has to travel even faster”. The sequence of events reveals the upheaval need to strengthen the ability of the early warning system, risk reduction, and management of national and global risks. Digital contact tracing apps like Aarogya setu (India) and Pan-European privacy preserving proximity tracing (German) has somehow helped but they are more effective in the initial stage and less relevant in the community spread phase. Thus, there is a need to devise a Decision Support System (DSS) based on machine learning algorithms. In this paper, we have attempted to propose an Additive Utility Assumption Approach for Criterion Comparison in Multi-criterion Intelligent Decision Support system for COVID-19. The dataset of Covid-19 has been taken from government link for validating the results. In this paper, an additive utility assumption-based approach for multi-criterion decision support system (MCDSS) with an accurate prediction of identified risk factors on certain well-defined input parameters is proposed and validated empirically using the standard SEIR model approach (Susceptible, Exposed, Infected and Recovered). The results includes comparative analysis in tabular form with already existing approaches to illustrate the potential of the proposed approach including the parameters such as Precision, Recall and F-Score. Other advanced parameters such as, MCC (Matthews Correlation Coefficient), ROC (Receiver Operating Characteristics) and PRC (Precision Recall) have also been considered for validation and the graphs are illustrated using Jupyter notebook. The statistical analysis of the most affected top eight states of India is undertaken effectively using the Weka software tool and IBM Cognos software to correctly predict the outbreak of pandemic situation due to Covid-19. Finally, the article has immense potential to contribute to the COVID-19 situation and may prove to be instrumental in propelling the research interest of researchers and providing some useful insights for the current pandemic situation., Graphical abstract, Highlights • Proposing a Multi-criterion Intelligent DSS • Epidemiology modeling of data using a SEIFR Model • Our approach applies data to get trained using multiple parameters to handle and serve significant challenges existing especially in healthcare.

Highlights • An effective cloud computing framework is proposed which predicts H1N1 infected patients and provides preventions to control infection rate. • Random decision tree is used to initially assess the infection in any patient depending on his/her symptoms. • Social Network Analysis (SNA) is used to present the state of outbreak. • Proposed framework is tested on synthetic data generated for two million users. • System provided 94% accuracy for classification and around 81% of resources utilization on Amazon EC2 cloud., H1N1 is an infectious virus which, when spread affects a large volume of the population. It is an airborne disease that spreads easily and has a high death rate. Development of healthcare support systems using cloud computing is emerging as an effective solution with the benefits of better quality of service, reduced costs and flexibility. In this paper, an effective cloud computing architecture is proposed which predicts H1N1 infected patients and provides preventions to control infection rate. It consists of four processing components along with secure cloud storage medical database. The random decision tree is used to initially assess the infection in any patient depending on his/her symptoms. Social Network Analysis (SNA) is used to present the state of the outbreak. The proposed architecture is tested on synthetic data generated for two million users. The system provided 94% accuracy for the classification and around 81% of the resource utilization on Amazon EC2 cloud. The key point of the paper is the use of SNA graphs to calculate role of an infected user in spreading the outbreak known as Outbreak Role Index (ORI). It will help government agencies and healthcare departments to present, analyze and prevent outbreak effectively.

In this paper, we propose a new way to carry out fitness evaluation in dynamic Particle Swarm Clustering (PSC) with centroid-based encoding. Generally, the PSC fitness function is selected among the clustering validity indices and most of them directly depend on the cluster centroids. In the traditional fitness evaluation approach, the cluster centroids are replaced by the centroids proposed by a particle position. We propose to first compute the centroids of the corresponding clusters and then use these computational centroids in fitness evaluation. The proposed way is called Fitness Evaluation with Computational Centroids (FECC). We conducted an extensive set of comparative evaluations and the results show that FECC leads to a clear improvement in clustering results compared to the traditional fitness evaluation approach with most of the fitness functions considered in this study. The proposed approach was found especially beneficial when underclustering is a problem. Furthermore, we evaluated 31 fitness functions based on 17 clustering validity indices using two PSC methods over a large number of synthetic and real data sets with varying properties. We used three different performance criteria to evaluate the clustering quality and found out that the top three fitness functions are Xu index, WB index, and Dunn variant DU 23 applied using FECC. These fitness functions consistently performed well for both PSC methods, for all data distributions, and according to all performance criteria. In all test cases, they were clearly among the better half of the fitness functions and, in the majority of the cases, they were among the top 4 functions. Further guidance for improved fitness function selection in different situations is provided in the paper.

We provide an introduction to Multiple Classifier Systems (MCS) including basic nomenclature and describing key elements: classifier dependencies, type of classifier outputs, aggregation procedures, architecture, and types of methods. This introduction complements other existing overviews of MCS, as here we also review the most prevalent theoretical framework for MCS and discuss theoretical developments related to MCS The introduction to MCS is then followed by a review of the application of MCS to the particular field of multimodal biometric person authentication in the last 25 years, as a prototypical area in which MCS has resulted in important achievements. This review includes general descriptions of successful MCS methods and architectures in order to facilitate the export of them to other information fusion problems. Based on the theory and framework introduced here, in the companion paper we then develop in more technical detail recent trends and developments in MCS from multimodal biometrics that incorporate context information in an adaptive way. These new MCS architectures exploit input quality measures and pattern-specific particularities that move apart from general population statistics, resulting in robust multimodal biometric systems. Similarly as in the present paper, methods in the companion paper are introduced in a general way so they can be applied to other information fusion problems as well. Finally, also in the companion paper, we discuss open challenges in biometrics and the role of MCS to advance them, This work was funded by projects CogniMetrics (TEC2015-70627-R) from MINECO/FEDER and RiskTrakc (JUST-2015-JCOO-AG-1). Part of thisthis work was conducted during a research visit of J.F. to Prof. Ludmila Kuncheva at Bangor University (UK) with STSM funding from COST CA16101 (MULTI-FORESEE)

Searching for scientific publications online is an essential task for researchers working on a certain topic. However, the extremely large amount of scientific publications found in the web turns the process of finding a publication into a very difficult task whereas, locating peers interested in collaborating on a specific topic or reviewing literature is even more challenging. In this paper, we propose a novel architecture to join multiple bibliographic sources, with the aim of identifying common research areas and potential collaboration networks, through a combination of ontologies, vocabularies, and Linked Data technologies for enriching a base data model. Furthermore, we implement a prototype to provide a centralized repository with bibliographic sources and to find similar knowledge areas using data mining techniques in the domain of Ecuadorian researchers community.

Internet of Things (IoT) has recently brought an influential research and analysis platform in a broad diversity of academic and industrial disciplines, particularly in healthcare. The IoT revolution is reshaping current healthcare practices by consolidating technological, economic, and social views. Since December 2019, the spreading of COVID-19 across the world has impacted the world’s economy. IoT technology integrated with Artificial Intelligence (AI) can help to address COVID-19. UAVs equipped with IoT devices can collect raw data that demands computing and analysis to make intelligent decision without human intervention. To mitigate the effect of COVID-19, in this paper, we propose an IoT-UAV-based scheme to collect raw data using onboard thermal sensors. The thermal image captured from the thermal camera is used to determine the potential people in the image (of the massive crowd in a city), which may have COVID-19, based on the temperature recorded. An efficient hybrid approach for a face recognition system is proposed to detect the people in the image having high body temperature from infrared images captured in a real-time scenario. Also, a face mask detection scheme is introduced, which detects whether a person has a mask on the face or not. The schemes’ performance evaluation is done using various machine learning and deep learning classifiers. We use the edge computing infrastructure (onboard sensors and actuators) for data processing to reduce the response time for real-time analytics and prediction. The proposed scheme has an average accuracy of 99.5% using various performance evaluation metrics indicating its practical applicability in real-time scenarios.

Forecasting and prediction of natural events, such as tropical and extra-tropical cyclones, inland flooding, and severe winter weather, provide critical guidance to emergency managers and decision-makers from the local to the national level, with the goal of minimizing both human and economic losses. This guidance is used to facilitate evacuation route planning, post-disaster response and resource deployment, and critical infrastructure protection and securing, and it must be available within a time window in which decision makers can take appropriate action. This latter element is that which induces the need for urgency in this area. In this paper, we outline the North Carolina Forecasting System (NCFS) for storm surge and waves for coastal North Carolina, which is threatened by tropical cyclones about once every three years. We initially used advanced cyberinfrastructure techniques (e.g., opportunistic grid computing) in an effort to provide timely guidance for storm surge and wave impacts. However, our experience has been that a distributed computing approach is not robust enough to consistently produce the real-time results that end users expect. As a result, our technical approach has shifted so that the reliable and timely delivery of forecast products has been guaranteed by provisioning dedicated computational resources as opposed to relying on opportunistic availability of external resources. Our experiences with this forecasting effort is discussed in this paper, with a focus on Hurricane Irene (2011) that impacted a substantial portion of the US east coast from North Carolina, up along the eastern seaboard, and into New England.

A computational imaging system based on wavefront coding is presented. Wavefront coding provides an extension of the depth-of-field at the expense of a slight reduction of image quality. This trade-off results from the amount of coding used. By using spatial light modulators, a flexible coding is achieved which permits it to be increased or decreased as needed. In this paper a computational method is proposed for evaluating the output of a wavefront coding imaging system equipped with a spatial light modulator, with the aim of thus making it possible to implement the most suitable coding strength for a given scene. This is achieved in an unsupervised manner, thus the whole system acts as a dynamically selfadaptable imaging system. The program presented here controls the spatial light modulator and the camera, and also processes the images in a synchronised way in order to implement the dynamic system in real time. A prototype of the system was implemented in the laboratory and illustrative examples of the performance are reported in this paper. Program summary Program title: DynWFC (Dynamic WaveFront Coding) Catalogue identifier: AEKC_v1_0 Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEKC_v1_0.html Program obtainable from: CPC Program Library, Queenʼs University, Belfast, N. Ireland Licensing provisions: Standard CPC licence, http://cpc.cs.qub.ac.uk/licence/licence.html No. of lines in distributed program, including test data, etc.: 10 483 No. of bytes in distributed program, including test data, etc.: 2 437 713 Distribution format: tar.gz Programming language: Labview 8.5 and NI Vision and MinGW C Compiler Computer: Tested on PC Intel ® Pentium ® Operating system: Tested on Windows XP Classification: 18 Nature of problem: The program implements an enhanced wavefront coding imaging system able to adapt the degree of coding to the requirements of a specific scene. The program controls the acquisition by a camera, the display of a spatial light modulator and the image processing operations synchronously. The spatial light modulator is used to implement the phase mask with flexibility given the trade-off between depth-of-field extension and image quality achieved. The action of the program is to evaluate the depth-of-field requirements of the specific scene and subsequently control the coding established by the spatial light modulator, in real time.

Highlights • A hierarchical temporal attention based model is proposed to support short-term and long-term human mobility sequence prediction. • The proposed hierarchical temporal attention incorporates individual mobility patterns into the model architecture. • The model is compared with four baseline methods on individual trajectory datasets with varying degree of traveling uncertainty. • Experiments demonstrate the outperformance of the proposed method using three evaluation metrics. • The proposed model uncovers individual frequential and periodical mobility patterns in an interpretable manner., Prediction of individual mobility is crucial in human mobility related applications. Whereas, existing research on individual mobility prediction mainly focuses on next location prediction and short-term dependencies between traveling locations. Long-term location sequence prediction is of great importance for long-time traffic planning and location advertising, and long-term dependencies exist as individual mobility regularity typically occurs daily and weekly. This paper proposes a novel hierarchical temporal attention-based LSTM encoder-decoder model for individual location sequence prediction. The proposed hierarchical attention mechanism captures both long-term and short-term dependencies underlying in individual longitudinal trajectories, and uncovers frequential and periodical mobility patterns in an interpretable manner by incorporating the calendar cycle of individual travel regularities into location prediction. More specifically, the hierarchical attention consists of local temporal attention to identify highly related locations in each day, and global temporal attention to discern important travel regularities over a week. Experiments on individual trajectory datasets with varying degree of traveling uncertainty demonstrate that our method outperforms four baseline methods on three evaluation metrics. In addition, we explore the interpretability of the proposed model in understanding individual daily, and weekly mobility patterns by visualizing the temporal attention weights and frequent traveling patterns associated with locations.

In biomedical domain, abbreviations are appearing more and more frequently in various data sets, which has caused significant obstacles to biomedical big data analysis. The dictionary-based approach has been adopted to process abbreviations, but it cannot handle ad hoc abbreviations, and it is impossible to cover all abbreviations. To overcome these drawbacks, this paper proposes an automatic abbreviation expansion method called LMAAE (Language Model-based Automatic Abbreviation Expansion). In this method, the abbreviation is firstly divided into blocks; then, expansion candidates are generated by restoring each block; and finally, the expansion candidates are filtered and clustered to acquire the final expansion result according to the language model and clustering method. Through restrict the abbreviation to prefix abbreviation, the search space of expansion is reduced sharply. And then, the search space is continuous reduced by restrained the effective and the length of the partition. In order to validate the effective of the method, two types of experiments are designed. For standard abbreviations, the expansion results include most of the expansion in dictionary. Therefore, it has a high precision. For ad hoc abbreviations, the precisions of schema matching, knowledge fusion are increased by using this method to handle the abbreviations. Although the recall for standard abbreviation needs to be improved, but this does not affect the good complement effect for the dictionary method., Highlights • An automatic method is proposed for biomedical big data. • This method overcomes the weaknesses of the traditional dictionary-based method. • An optimization strategy is proposed for time complexity improvement. • An excessive number of expansions causes poor abbreviation sense disambiguation.

Today, although there is an increasing interest in temporal JSON instance documents, since they allow tracking data changes, recovering past data versions, and executing temporal queries, there is no support (data model, modelling language, method, or tool) for conceptual modelling of temporal JSON data. Moreover, even though there are some graphical editors to build JSON Schemata (like JSON Schema Editor of Altova), they do not provide any built-in support for modelling temporal aspects of JSON data. Therefore, designers of JSON-based NoSQL data stores are proceeding in an ad hoc manner when they have to model some temporal requirements. To fill this theoretical and practical gap, we propose in this paper a graphical conceptual model for time-varying JSON data, named Temporal JSON Conceptual Model (TempoJCM). To this purpose, first we define a graphical conceptual model for conventional (i.e., non temporal) JSON data, called JSON Conceptual Model (JCM), and then we extend it to support modelling of temporal aspects of JSON data. TempoJCM facilitates conceptual modelling of both conventional and temporal JSON data, in a graphical and user-friendly manner. An editor supporting TempoJCM is planned to become the user interface for temporal JSON schema design in the τJSchema framework.

Traditional explicit authentication mechanisms, in which the device remains unlocked after the introduction of some kind of password, are slowly being complemented with the so-called implicit or continuous authentication mechanisms. In the latter, the user is constantly monitored in one or more ways, in search for signs of unauthorized access, which may happen if a third party has access to the phone after it has been unlocked. There are some different forms of continuous authentication, some of which based on Machine Learning. These are generally black box models, that provide a decision but not an explanation. In this paper we propose an approach for continuous authentication based on behavioral biometrics, machine learning, and that includes domain-dependent aspects for the user to interpret the actions and decisions of the system. It is non-intrusive, does not require any additional hardware, and can be used continuously to monitor user identity., This work has been supported by national funds through FCT – Fundação para a Ciência e Tecnologia through project UIDB/04728/2020.

This paper presents a local descriptor coding scheme for multicamera surveillance and 3D reconstruction embedding an autoencoder into a traditional distributed source coding strategy. The proposed solution permits shifting most of the computational complexity at the decoder/receiver and exploiting the correlation among descriptors of different cameras (thus reducing the coded bit rate) without increasing the inter-device communication load. Experimental results show that the proposed scheme permits obtaining a satisfying accuracy with respect to the most recent solutions while generating a limited bit rate.

Network failures can significantly degrade the network performance in several ways, ranging from experiencing unacceptable delays (e.g., in video conference applications) to leading to significant drops in sales for companies (e.g., airlines tickets purchases). In this paper, we focus on network failures caused by links that are prone to fail due to their advanced stage of lifetime. In particular, the formalization of the Hardware Deterioration Failure Mode (HDFM) for network line cards is provided to evaluate the penalty for network operators of handling HDFM failing links using classical methods based on Bidirectional Forwarding Detection and Fast Re-Route mechanisms. An algorithm, named Hardware Deterioration Detection (HDD), is proposed to detect if links are experiencing failures as well as to provide an estimation of their remaining lifetime. The performance evaluation over data generated by our provided network failure simulator shows that HDD is able to correctly classify links as failing or not failing, regardless of the type of failure considered. Moreover, overall descriptions for failure models in which HDD is able to produce high quality estimations of the remaining lifetime have been identified, with a trade-off between the hardware deterioration time dependence and the relationship between lifetime and packet loss probability.

The advancements in networking technologies have introduced the Internet of Everything (IoE) and smart living concepts. The main idea behind making everything smarter is to improve individuals’ quality of life. An excellent example of such a technology is smart healthcare which provides efficient, sustainable, and real-time human services. However, data security and privacy are among the most important challenges of smart healthcare applications. Blockchain (BC) has been considered as a promising solution for the secure management of healthcare data due to its immutability and transparency features. However, there is a trade-off between transparency and user data privacy which is a prominent challenge in adopting BC for healthcare applications. Some researchers have considered user data privacy and proposed a few solutions; however, data owner’s access control desire has not been considered in the state-of-the-art models. In this paper, to overcome the trade-off challenge between transparency and access control, we propose an architecture (so-called BCHealth) that enables data owners to define their desired access policies over their privacy-sensitive healthcare data. BCHealth is composed of two separate chains for storing access policies and data transactions. We address the real-world development challenges of BC, i.e., scalability, delay, and overhead, by adopting a clustering approach. Our extensive experimental analysis proves the efficiency of BCHealth (in terms of computation and processing time) and its resilience against several security attacks.

In this paper we propose a robust fuzzy clustering model, the STAR-based Fuzzy C-Medoids Clustering model with Noise Cluster, to define territorial partitions of the European regions (NUTS2) according to the workplaces mobility trends for places of work provided by Google with reference to the whole COVID-19 pandemic period. The clustering model takes into account both temporal and spatial information by means of the autoregressive temporal and spatial coefficients of the STAR model. The proposed clustering model through the noise cluster is capable of neutralizing the negative effects of noisy data. The main empirical results regard the expected direct relationship between the Community mobility trend and the lockdown periods, and a clear spatial interaction effect among neighboring regions.

In this paper, we evaluate the performance of machine learning classifiers (Logistic Regression, CART, Random Forest) by fabricating adversarial examples (malware samples) statistically identical to goodware. To this end, we demonstrate three scenarios, (a) random attribute injection (b) insertion of prominent attributes from legitimate apps and (c) poisoning of class labels, for creating tainted malware samples, to mislead reduce accuracy of classification models. Experiments were conducted on data-set consisting of 15649 android applications comprising 5373 malicious and 10276 legitimate apps. The outcome of investigations demonstrates significant drop in accuracies in the range of 12-50%. However, in the absence of adversarial examples in the test set, the performance of classifiers was observed between 94.8-97.9%.

The detection mechanism provided by current antimalware is the so-called signature based, requiring that a threat must be widespread to be recognised by the antimalware. Even if a malware is rightly recognized, by applying even trivial obfuscation techniques, it is really easy to bypass the antimalware detection mechanism. In this paper we propose a method to detect if an Android application is obfuscated with the call indirection obfuscation techniques by exploiting formal equivalence checking. In the experimental analysis we show the effectiveness of the propose approach for call indirection obfuscation technique detection, by exploiting two obfuscation tools.

OWL 2 QL is a standard profile of the OWL 2 ontology language, specifically tailored to Ontology-Based Data Management. Inspired by recent work on higher-order Description Logics, in this paper we present a new semantics for OWL 2 QL ontologies, called Metamodeling Semantics (MS), and show that, in contrast to the official Direct Semantics (DS) for OWL 2 , it allows exploiting the metamodeling capabilities natively offered by the OWL 2 punning. We then extend unions of conjunctive queries with both metavariables, and the possibility of using TBox atoms, with the purpose of expressing meaningful metalevel queries. We first show that under MS both satisfiability checking and answering queries including only ABox atoms, have the same complexity as under DS. Second, we investigate the problem of answering general metaqueries, and single out a new source of complexity coming from the combined presence of a specific type of incompleteness in the ontology, and of TBox axioms among the query atoms. Then we focus on a specific class of ontologies, called TBox-complete, where there is no incompleteness in the TBox axioms, and show that general metaquery answering in this case has again the same complexity as under DS. Finally, we move to general ontologies and show that answering general metaqueries is coNP-complete with respect to ontology complexity, Π 2 p -complete with respect to combined complexity, and remains AC 0 with respect to ABox complexity.

A new version of the code system pengeom , which provides a complete set of tools to handle different geometries in Monte Carlo simulations of radiation transport, is presented. The distribution package consists of a set of Fortran subroutines and a Java graphical user interface that allows building and debugging the geometry-definition file, and producing images of the geometry in two- and three-dimensions. A detailed description of these tools is given in the original paper [ Comput. Phys. Commun. 199 (2016) 102–113] and in the code manual included in the distribution package. The present new version corrects a bug in the Fortran subroutines, and it includes various improvements of the Java graphical user interface. New Version Program Summary Program Title: pengeom CPC Library link to program files: https://doi.org/10.17632/zgswr8kyf5.1 Licensing provisions: CC BY NC 3.0 Programming language: Fortran 90, Java Journal reference of previous version: Comput. Phys. Commun. 199 (2016) 102–113 Does the new version supersede the previous version?: Yes Reasons for the new version: The original Fortran code contained a subtle bug, which had an effect only when (1) a particle moving within a void region entered a material body, and (2) the particle was left in the void volume because of numerical roundoff errors (undershot). In such case, the steering main program erroneously concluded that the particle was in the external vacuum and, consequently, it discontinued the simulation of the particle trajectory. This anomalous behavior has been corrected by shifting the position of the particle a small distance, determined by the parameters of the master ray equation, which places the particle a small distance inside the material body. In the original Fortran subroutines, inconsistencies in the geometry-definition file resulted in a stop of the program; this reaction was hard to control from a graphical user interface. The present new version has been verified to work correctly under Java 8, Java 11, and Java 13. Summary of revisions: In the new version the tracking bug is corrected. In addition, any apparent syntax error is now resolved by returning the control to the calling program with an error flag activated and with an error message. Various aspects of the Java user interface have been improved: the text editor now includes line numbers, which are used to locate possible inconsistencies and errors in the geometry definition file; the size of the application window can now be modified, and the possible sizes of the graphics windows are adapted to the actual resolution of the monitor; the action of keystroke commands under Windows, Linux, and macOS is now the same. Nature of problem: The Fortran subroutines perform all geometry operations in Monte Carlo simulations of radiation transport with arbitrary interaction models. They track particles through complex quadric geometries, i.e., through material systems consisting of homogeneous bodies limited by quadric surfaces. Particles are moved in steps (free flights) of a given length, which is dictated by the simulation program, and are halted when they cross an interface between media of different compositions or when they enter selected bodies. At the end of each step, pengeom returns the indices of the material and body where the particle is. Solution method: pengeom is tailored to optimize simulation speed and accuracy. Fast tracking is accomplished by the use of quadric surfaces, which facilitate the calculation of ray intersections , and of modules (connected volumes limited by quadric surfaces) organized in a hierarchical structure. Optimal accuracy is obtained by considering fuzzy surfaces, with the aid of a simple algorithm that keeps control of multiple intersections of a ray and a surface. The 64-bit Java GUI PenGeomJar provides a complete geometry toolbox; it allows building and debugging of the geometry definition file, as well as visualizing the resulting geometry in two and three dimensions. Additional comments including restrictions and unusual features: All geometrical operations are performed internally. The connection between the steering main program and the tracking routines is through a Fortran module, which contains the state variables of the transported particle, and the input-output arguments of subroutine STEP . Rendering of two- and three-dimensional images is performed by using the pengeoem subroutines, so that what we see in the images is what is passed to the simulation program. By default pengeom can handle systems with up to 5,000 bodies and 10,000 surfaces. These numbers can be increased by editing the source file. [1] J. Almansa, F. Salvat-Pujol, G. Diaz-Londono, A. Carnicer, A.M. Lallena, and Francesc Salvat, Comput. Phys. Commun. 199 (2016) 102–113.

In modern manufacturing, each stage of industrial processes is accurately measured via multiple sensors and, consequently, a large amount of data is made available for analytics, monitoring and control purposes. A possible use of such data is to detect anomalies in order to prevent potential damages and hazards. In this paper, we will consider a sensor setup returning distributed time series measurements that can be used for failure identification. In particular, an anomaly detection strategy based on Vector Autoregressive (VAR) modeling for multivariate time series will be presented and analyzed in detail. The effectiveness of the proposed methodology will be assessed on experimental data from a real industrial case study.

Unmanned Aerial Vehicles (UAV) have revolution- ized the aircraft industry in this decade. UAVs are now capable of carrying out remote sensing, remote monitoring, courier delivery, and a lot more. A lot of research is happening on making UAVs more robust using energy harvesting techniques to have a better battery lifetime, network performance and to secure against attackers. UAV networks are many times used for unmanned missions. There have been many attacks on civilian, military, and industrial targets that were carried out using remotely controlled or automated UAVs. This continued misuse has led to research in preventing unauthorized UAVs from causing damage to life and property. In this paper, we present a literature review of UAVs, UAV attacks, and their prevention using anti-UAV techniques. We first discuss the different types of UAVs, the regulatory laws for UAV activities, their use cases, recreational, and military UAV incidents. After understanding their operation, various techniques for monitoring and preventing UAV attacks are described along with case studies.

The voluminous amount of data generated by individuals’ mobile sensors and wearable devices is considered of a great value for the benefits of patients and clinical research. Recent advances incorporating data mining and cloud computing have leveraged the great potential of these data. However, the introduction of such technologies in the process of mobile crowd sensed data mining and analytics could potentially lead to security and privacy concerns. Individuals and organizations are not able to share and collectively run computations on their private data captured by different sensors to infer any processes of common interest. Although solutions such as Secure Multiparty Computation (SMC) were laid decades ago, they are still perceived for theoretical interest only, so far. In this paper, we aim at bridging the gap between privacy-preserving data mining and its practice. To do so, we introduce a blockchain-based privacy-preserving SVM classification (BPPSVC) between mutually distrustful data owners. In BPPSVC, blockchain technology along with smart contracts underlay more realistic assumptions about the adversarial model. Our main focus is on investigating the immutability, security and the bookkeeping properties of the blockchain in preserving the privacy of an SVM classifier over horizontally distributed IoT data. To this end, we first propose the system architecture, adversary model and design goals of BPPSVC, then we describe the design details. Our security analysis indicates that the proposed system is secure and it provides fairness and protection against Denial of Service (DoS) attacks. We finally show the efficiency and feasibility of BPPSVC through rigorous experimental results. This work was partially supported by the Program of International Science and Technology Cooperation and Exchange of Sichuan Province, China (2017HH0028, 2018HH0102, 2019YFH0014, 2020YFH0030), and by the Science and Technology Program of Sichuan Province, China (2020YFSY0061). This work was also sponsored by CCF-Tencent Open Research Fund WeBank Special Funding, China.

Identifying cyber attacks traffic is very important for the Internet of things (IoT) security in smart city. Recently, the research community in the field of IoT Security endeavor hard to build anomaly, intrusion and cyber attacks traffic identification model using Machine Learning (ML) algorithms for IoT security analysis. However, the critical and significant problem still not studied in depth that is how to select an effective ML algorithm when there are numbers of ML algorithms for cyber attacks detection system for IoT security. In this paper, we proposed a new framework model and a hybrid algorithm to solve this problem. Firstly BoT-IoT identification dataset is applied and its 44 effective features are selected from a number of features for the machine learning algorithm. Then five effective machine learning algorithm is selected for the identification of malicious and anomaly traffic identification and also select the most widely ML algorithm performance evaluation metrics. To find out which ML algorithm is effective and should be used to select for IoT anomaly and intrusion traffic identification, a bijective soft set approach and its algorithm is applied. Then we applied the proposed algorithm based on bijective soft set approach. Our experimental results show that the proposed model with the algorithm is effective for the selection ML algorithm out of numbers of ML algorithms. This work is supported by the National Key research and Development Plan (Grant No. 2018YFB0803504), the Guangdong Province Key Research and Development Plan (Grant No.2019B010137004) and the National Natural Science Foundation of China under Grant No. 61871140, and Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (2019).

Computer system vulnerabilities, computer viruses, and cyber attacks are rooted in software vulnerabilities. Reducing software defects, improving software reliability and security are urgent problems in the development of software. The core content is the discovery and location of software vulnerability. However, traditional human experts-based approaches are labor-consuming and time-consuming. Thus, some automatic detection approaches are proposed to solve the problem. But, they have a high false negative rate. In this paper, a deep learning based static taint analysis approach is proposed to automatically locate Internet of Things (IoT) software vulnerability, which can relieve tedious manual analysis and improve detection accuracy. Deep learning is used to detect vulnerability since it considers the program context. Firstly, the taint from the difference file between the source program and its patched program selection rules are designed. Secondly, the taint propagation paths are got using static taint analysis. Finally, the detection model based on two-stage Bidirectional Long Short Term Memory (BLSTM) is applied to discover and locate software vulnerabilities. The Code Gadget Database is used to evaluate the proposed approach, which includes two types of vulnerabilities in C/C++ programs, buffer error vulnerability (CWE-119) and resource management error vulnerability (CWE-399). Experimental results show that our proposed approach can achieve an accuracy of 0.9732 for CWE-119 and 0.9721 for CWE-399, which is higher than that of the other three models (the accuracy of RNN, LSTM, and BLSTM is under than 0.97) and achieve a lower false negative rate and false positive rate than the other approaches. This work was supported in part by the National Key R&D Plan under Grant CNS 2016QY06X1205 ,in part by the Basic Research Business Fees of Central Colleges under Grant CNS 20826041B4252 , in part by the National Natural Science Foundation (NSFC) under Grant CNS 61572115 , and in part by the Science and Technology Project of State Grid Corporation of China under Grant CNS 522722180007 . Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of the funding agencies.

In the Industry 4.0 vision, Prognostics and Health Management (PHM) is expected to assist domain experts in the generation of maintenance decisions. PHM relies on the processing of data sensed from the manufacturing plant for inferring the future performance of production systems. The acquisition and management of data brings different challenges such as data integration, heterogeneity, search usability and volume. To be best of authors’ knowledge, an information system for sharing maintenance data able to fulfill the aforementioned challenges is not available yet. XRepo is proposed within this paper and faces three of the identified challenges through selected functionality: i) Heterogeneity: stored data is complaint with a standard format that includes the information necessary for performing PHM analysis; ii) Integration: data is uploaded to the repository through files or web services; iii) Search usability: stored data can be filtered by criteria and downloaded. This work is meant to be an initial effort towards the generation of a common information system for PHM analysis.

Context: Software architecture viewpoints modularize the software architectures in terms of different viewpoints that each address a different concern. Unified Modeling Language (UML) is so popular among practitioners for modeling software architectures from different viewpoints. Objective: In this paper, we aimed at understanding the practitioners’ UML usage for the modeling of software architectures from different viewpoints. Method: To this end, 109 practitioners with diverse profiles have been surveyed to understand practitioners’ UML usage for six different viewpoints: functional, information, concurrency, development, deployment, and operational. Each viewpoint has been considered in terms of a set of software models that can be created in that viewpoint. Results: The survey includes 35 questions for different viewpoint models, and the results lead to interesting findings. While the top popular viewpoints for the UML-based software architecture modeling are the functional (96%) and information (99%) viewpoints, the least popular one is the operational viewpoint that is considered by 26% of the practitioners. The top popular UML modeling tool is Enterprise Architect regardless of the viewpoints considered. Concerning the software models that can be created in each viewpoint, UML's class diagram is practitioners’ top choice for the functional structure (71%), data structure (85%), concurrency structure (75%), software code structure (34%), and system installation (39%), and system support (16%) models; UML's sequence diagram is the top choice for the data lifecycle models (47%); UML's deployment diagram for the physical structure (71%), mapping between the functional and physical components (53%), and system migration (21%) models; UML's activity diagram for the data flow (65%), software build and release processes (20–22%), and system administration (36%) models; UML's component diagram for the mapping between the functional and concurrent components (35%), software module structure (47%), and system configuration (21%) models; and UML's package diagram for the software module structure (47%) models. © 2020 Elsevier B.V.

A circular economy is an industrial system that is restorative or regenerative by intention or design. During the last decade, the circular economy became an attractive paradigm to increase global welfare while minimizing the environmental impacts of economic activities. Although several studies concerning the potential benefits and drawbacks of policies that implement the new paradigm have been performed, there is currently no standardized theoretical model or software to execute such assessment. In order to fill this gap, in the present paper we show how to perform these analyses using Environmentally Extended Input-Output Analysis. We also describe a python package (pycirk) for modeling Circular Economy scenarios in the context of the Environmentally Extended Multi-Regional Input-Output database EXIOBASE V3.3, for the year 2011. We exemplify the methods and software through a what-if zero-cost case study on two circular economy strategies (Resource Efficiency and Product Lifetime Extension), four environmental pressures and two socio-economic factors. © 2019 The Authors

In this paper, a method for generating a mixed integer linear programming problem from deterministic timed Petri nets to optimize the makespan of manufacturing systems is proposed. After exposing several challenges involved in such a model transformation, a novel method is exploited to overcome them. The adopted solution may create a synergy between the highly expressive timed Petri net modelling framework and the computational attractive mathematical programming tool set.

Functionally Graded Materials (FGM) are a class of advanced composites with a gradual and continuously varying microstructural composition. This paper presents MicroFEA 1.0, which is a software package developed to help overcoming the challenges of designing and analyzing FGM structures. It provides capabilities to fit FGM experimental data to B-Spline curves and to analyze composite materials using mean-field micromechanics homogenization techniques. The package is composed of MATLAB scripts and Fortran subroutines intended for Finite Element Analysis (FEA) using Abaqus. The Abaqus user-material subroutines (UMATs) are designed to work with heterogeneous materials at the integration points level, discarding the need for custom-designed elements. Thermomechanical analyses can also be carried out with the help of UTEMPs. This software will provide valuable help for scientists and engineers facing the development of new FGM structural components.

Network Function Virtualization enables network slicing as a novel paradigmfor service provisioning. With network slicing, Virtual Network Functions(VNFs) can be instantiated at different locations of the infrastructure, choosing their optimal placement based on parameters such as the requirementsof the service or the resources available. One limitation of state-of-the-arttechnology for network slicing is the inability to re-evaluate orchestration decisions once the slice has been deployed, in case of changing service demandsor network conditions.In this paper, we present ACHO, a novel software framework that enablesseamless re-orchestration of VNFs of any kind, including RAN and Core.With ACHO, VNFs and resources can be easily re-assigned to match, e.g.,varying user demands or changes in the nodes' load. ACHO uses lightweightmechanisms, such as splitting the engine of a VNF from the data it requiresto perform its operation, in such a way that, when re-allocating a VNF, onlythe data is moved (a new engine is instantiated in the new location). Wedemonstrate the use of ACHO in a small scale testbed, showing that (i) theproposed re-orchestration is feasible, (ii) it results much faster than existingalternatives (especially for relocation), and (iii) the framework can be readilyapplied to existing VNFs after minimal changes to their implementation. This work was partially supported by the European Commission in the framework of the H2020 5G-PPP 5G-TOURS (Grant no. 856950) project and the 5G-EVE (Grant no. 815074) project.

In this paper the Power-Oriented Graphs (POG) technique is introduced and a new modeling program named “POG Modeler”, freely available on the web, is presented. In the POG Modeler program the physical systems can be defined graphically using an ascii commend-line interface and referring to predefined graphic symbols. The POG Modeler automatically analyzes the given physical system and provides the following outputs: 1) the differential equations of the given system in symbolic form: 2) The POG block scheme of the considered system; 3) the Simulink block scheme of the given system ready for the Matlab environment. The POG systems are simple block schemes that can be easily used also by beginners.

Neural attention has become a key component in many deep learning applications, ranging from machine translation to time series forecasting. While many variations of attention have been developed over recent years, all share a common component in the application of a softmax function to normalize the attention weights, in order to transform them into valid mixing coefficients. In this paper, we aim to improve the modeling flexibility of a generic attention module by innovatively replacing this softmax operation with a learnable softmax, in which the normalizing functions are also adapted from the data. Specifically, our generalized softmax builds upon recent work in learning activation functions for deep networks, in particular the kernel activation function and its extensions. We describe the application of the proposed technique for the challenging case of time series forecasting with the dual-stage attention-based recurrent neural network (DA-RNN), an innovative model for predicting time series that employs two different attention modules for handling exogenous factors and long-term dependencies. A series of real-world benchmarks are used to show that simply plugging-in our generalized attention model can improve results on all datasets, even when keeping the number of trainable parameters in the model constant. To further evaluate the algorithm, we collect a novel dataset for predicting the Bitcoin closing exchange rate, a problem of high practical significance lately. Finally, to foster research in the topic, we also release both the dataset and our model as an open source extensible library. Over a baseline DA-RNN, our proposed model delivers an improvement of MAR ranging from 6% to 15% using our newly-released dataset.

The problem of vehicle autonomous driving currently represents a topic of great interest from both theoretical and practical points of view. Among the challenging tasks to be addressed within any autonomous driving framework, one of the most important ones is localization from data collected in real time. Within such framework, this paper is specifically focused on the localization problem for rail vehicles, such as railway and tramway vehicles. Our specific interest is on investigating solutions to the localization problem which are (as much as possible) independent on ground sensor infrastructure and are therefore suitable to be employed on any rail vehicle, irrespective of the ground equipment of the specific tracks. To this end, we refer to a multi-sensor framework and, specifically, to a sensor fusion scheme which collects data from different sensors installed on the vehicle (namely, an Inertial Measurement Unit and a Global Positioning System) and carries out a Kalman-based filtering recursion which relies on a simplified vehicle model. With the aim of identifying a solution for the localization problem providing desirable performance, we carry out a comparative simulation analysis concerning different Kalman-based data fusion strategies (in particular, the Extended Kalman Filter and the Unscented Kalman Filter are considered).

In this paper, we propose a suboptimal distributed LQR control method, applicable to systems coupled through both physical interconnections and the quadratic cost to be minimized. Thanks to a novel suboptimal but distributed cost-to-go matrix update that enforces block-diagonality, the suboptimal LQR gain matrix is structured, making the overall control scheme distributed. Moreover, the proposed control design algorithm is scalable. Theoretical properties of the method, including the stability of the closed-loop system, are investigated. A case study is shown to illustrate the features of the approach. Copyright (C) 2020 The Authors.

Radiological reports are a valuable source of textual information, which can be exploited to improve clinical care and to support research. Such information can be extracted and put into a structured form using machine learning techniques. Some of them rely not only on the classification labels but also on the manual annotation of relevant snippets, which is a time consuming job and requires domain experts. In this paper, we apply deep learning techniques and in particular Long Short Term Memory (LSTM) networks to perform such a task relying only on the classification labels. We focus on the classification of chest computed tomography reports in Italian according to a classification schema proposed for this task by the radiologists of Spedali Civili di Brescia. Each report is classified according to such schema using a combination of neural network classifiers. The resulting system is a novel classification system, which we compare to a previous system based on standard machine learning techniques which used annotations of relevant snippets.

Additive Manufacturing (AM) is already becoming part of our life from a technological, economic and social point of view. Nowadays, it is applied in several manufacturing sectors. In particular, AM shows huge opportunities in the medical field and for healthcare applications. Due to its capability to produce complex geometries directly working on medical 3D images and thanks to the possibility to 3D-print biocompatible materials, AM is a key technology for the fabrication both of external and internal medical devices. In particular, the use of AM for medical applications is typically articulated in three steps: 3D-scanning of the patient anatomy, segmentation the medical scan and elaboration through CAD software for the preparation of a STL file suitable for the AM process. One of the main research topic in this field is the definition and optimization of procedures that, taking precise data from an individual patient, could be applied to the design and fabrication of customized components for medical applications. Therefore, this paper presents a project aimed at the fabrication of customized tracheal stents starting from the actual patient anatomy. In particular, it follows an approach based on molds FDM fabrication followed by biocompatible silicone casting. Molds were designed to obtain a tracheal stent based the patient anatomical tracheal lumen and were fabricated using FDM technology. Moreover, since the surface roughness is one of the most critical aspects related to the FDM, the produced molds were finished with a chemical surface post-treatment based on the use of acetone vapours. Overall, the whole developed procedure results in an effective custom-made medical devices realization.

Within the area of intelligent User Interfaces, we propose what we call Sentient Embodied Conversational Agents (SECAs): virtual characters able to engage users in complex conversations and to incorporate sentient capabilities similar to the ones humans have. This paper introduces SECAs together with their architecture and a publicly available software library that facilitates their inclusion in applications –such as educational and elder-care– requiring proactive and sensitive agent behaviours. In fact, we illustrate our proposal with a virtual tutor embedded in an educational application for children. The evaluation was performed in two stages: firstly, we tested a version with basic textual processing capabilities; and secondly, we evaluated a SECA with Machine-Learning-enhanced user understanding capabilities. The results show a significant improvement in users’ perception of the agent’s understanding capability. Indeed, the Response Error Rate decreased from 22.31% to 11.46% using ML techniques. Moreover, 99.33% of the participants consider the global experience of talking with the virtual tutor with sentient capabilities to be satisfactory.

Deep Neural networks have gained lots of attention in recent years thanks to the breakthroughs obtained in the field of Computer Vision. However, despite their popularity, it has been shown that they provide limited robustness in their predictions. In particular, it is possible to synthesise small adversarial perturbations that imperceptibly modify a correctly classified input data, making the network confidently misclassify it. This has led to a plethora of different methods to try to improve robustness or detect the presence of these perturbations. In this paper, we perform an analysis of β-Variational Classifiers, a particular class of methods that not only solve a specific classification task, but also provide a generative component that is able to generate new samples from the input distribution. More in details, we study their robustness and detection capabilities, together with some novel insights on the generative part of the model.

Surveillance security cameras are increasingly deployed in almost every location for monitoring purposes, including watching people and their actions for security purposes. For criminology, images collected from these cameras are usually used after an incident occurs to analyze who could be the people involved. While this usage of the cameras is important for a post crime action, there exists the need for real time monitoring to act as an early warning to prevent or avoid an incident before it occurs. In this paper, we describe the development and implementation of an early warning system that recognizes people automatically in a surveillance camera environment and then use data from various sources to identify these people and build their profile and network. The current literature is still missing a complete workflow from identifying people/criminals from a video surveillance to building a criminal information extraction framework and identifying those people and their interactions with others We train a feature extraction model for face recognition using convolutional neural networks to get a good recognition rate on the Chokepoint dataset collected using surveillance cameras. The system also provides the function to record people appearance in a location, such that unknown people passing through a scene excessive number of times (above a threshold decided by a security expert) will then be further analyzed to collect information about them. We implemented a queue based system to record people entrance. We try to avoid missing relevant individuals passing through as in some cases it is not possible to add every passing person to the queue which is maintained using some cache handling techniques. We collect and analyze information about unknown people by comparing their images from the cameras to a list of social media profiles collected from Facebook and intelligent services archives. After locating the profile of a person, traditional news and other social media platforms are crawled to collect and analyze more information about the identified person. The analyzed information is then presented to the analyst where a list of keywords and verb phrases are shown. We also construct the person’s network from individuals mentioned with him/her in the text. Further analysis will allow security experts to mark this person as a suspect or safe. This work shows that building a complete early warning system is feasible to tackle and identify criminals so that authorities can take the required actions on the spot.

Coronavirus disease 2019 (COVID-19) is a world-wide epidemic and efficient prevention and control of this disease has become the focus of global scientific communities. In this paper, a novel face mask detection framework FMD-Yolo is proposed to monitor whether people wear masks in a right way in public, which is an effective way to block the virus transmission. In particular, feature extractor employs Im-Res2Net-101 which combines Res2Net module and deep residual network, where utilization of hierarchical convolutional structure, deformable convolution and non-local mechanisms enables thorough information extraction from the input. Afterwards, an enhanced path aggregation network En-PAN is applied for feature fusion, where high-level semantic information and low-level details are sufficiently merged so that the model robustness and generalization ability can be enhanced. Moreover, localization loss is designed and adopted in model training phase, and Matrix NMS method is used in the inference stage to improve the detection efficiency and accuracy. Benchmark evaluation is performed on two public databases with the results compared with other eight state-of-the-art detection algorithms. At IoU = 0.5 level, proposed FMD-Yolo has achieved the best precision AP50 of 92.0% and 88.4% on the two datasets, and AP75 at IoU = 0.75 has improved 5.5% and 3.9% respectively compared with the second one, which demonstrates the superiority of FMD-Yolo in face mask detection with both theoretical values and practical significance.

Made available in DSpace on 2019-10-05T23:26:50Z (GMT). No. of bitstreams: 0 Previous issue date: 2019-06-01 Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Concept drift methods learn patterns in non-stationary environments. Although such behavior is usually not expected in traditional classification problems, in real-world scenarios one can face them very much easier. In such a context, classifiers can be fooled and their effectiveness affected as well. Some examples include theft detection in energy distribution systems, where the consumer's behavior may change suddenly or smoothly, or even churn prediction in mobile companies. In this paper, we introduce the Optimum-Path Forest (OPF) classifier in the context of concept drift, using decisions for concept drift handling based on a committee of OPF classifiers. We consider three distinct perspectives (three rounds of experiments with variations of streaming managements) over publics datasets, being the results compared to the ones obtained by standard OPF. We consider OPF ensemble suitable to work under these dynamic scenarios since its recognition rates were considerably better when compared to traditional OPF. (C) 2019 Elsevier B.V. All rights reserved. Univ Fed Sao Carlos, Sao Paulo, Brazil Sao Paulo State Univ, Sao Paulo, Brazil Univ Fortaleza, Fortaleza, Ceara, Brazil Sao Paulo State Univ, Sao Paulo, Brazil FAPESP: 2013/07375-0 FAPESP: 2014/16250-9 FAPESP: 2014/12236-1 FAPESP: 2016/19200-8 CNPq: 306166/2014-3 CNPq: 304315/2017-6 CNPq: 307066/2017-7 CNPq: 430274/2018-1

Structural learning of Bayesian Networks (BNs) is a NP-hard problem, which is further complicated by many theoretical issues, such as the I-equivalence among different structures. In this work, we focus on a specific subclass of BNs, named Suppes-Bayes Causal Networks (SBCNs), which include specific structural constraints based on Suppes’ probabilistic causation to efficiently model cumulative phenomena. Here we compare the performance, via extensive simulations, of various state-of-the-art search strategies, such as local search techniques and Genetic Algorithms, as well as of distinct regularization methods. The assessment is performed on a large number of simulated datasets from topologies with distinct levels of complexity, various sample size and different rates of errors in the data. Among the main results, we show that the introduction of Suppes’ constraints dramatically improve the inference accuracy, by reducing the solution space and providing a temporal ordering on the variables. We also report on trade-offs among different search techniques that can be efficiently employed in distinct experimental settings. This manuscript is an extended version of the paper “Structural Learning of Probabilistic Graphical Models of Cumulative Phenomena” presented at the 2018 International Conference on Computational Science [1] .

In Marconi et al. 2007, it was shown that a solution to the output regulation problem for minimum-phase normal forms always exists. That approach, however, yields only an existence result, and no general analytic procedure is known to actually choose the regulator’s degrees of freedom, even for simple problems. In this paper we propose an adaptive regulator that, leveraging the aforementioned existence result, self-tunes online according to an optimization policy. To this aim, the regulator may employ every system identification scheme that fulfills some given strong stability properties, and the asymptotic regulation error is proved to be directly related to the prediction capabilities of the identifier.