Catálogo de publicaciones - revistas

<jats:p>Deep neural networks have proven to be particularly effective in visual and audio recognition tasks. Existing models tend to be computationally expensive and memory intensive, however, and so methods for hardware-oriented approximation have become a hot topic. Research has shown that custom hardware-based neural network accelerators can surpass their general-purpose processor equivalents in terms of both throughput and energy efficiency. Application-tailored accelerators, when co-designed with approximation-based network training methods, transform large, dense, and computationally expensive networks into small, sparse, and hardware-efficient alternatives, increasing the feasibility of network deployment. In this article, we provide a comprehensive evaluation of approximation methods for high-performance network inference along with in-depth discussion of their effectiveness for custom hardware implementation. We also include proposals for future research based on a thorough analysis of current trends. This article represents the first survey providing detailed comparisons of custom hardware accelerators featuring approximation for both convolutional and recurrent neural networks, through which we hope to inspire exciting new developments in the field.</jats:p>

<jats:p>Negative sequential patterns (NSPs) produced by negative sequence analysis (NSA) capture more informative and actionable knowledge than classic positive sequential patterns (PSPs) due to involving both occurring and nonoccurring items, which appear in many applications. However, the research on NSA is still at an early stage, and NSP mining involves very high computational complexity and a very large search space, there is no widely accepted problem statement on NSP mining, and different settings on constraints and negative containment have been proposed in existing work. Among existing NSP mining algorithms, there are no general and systemic evaluation criteria available to assess them comprehensively. This article conducts a comprehensive technical review of existing NSA research. We explore and formalize a generic problem statement of NSA; investigate, compare, and consolidate the definitions of constraints and negative containment; and compare the working mechanisms and efficiency of existing NSP mining algorithms. The review is concluded by discussing new research opportunities in NSA.</jats:p>

<jats:p>Insider threats are one of today’s most challenging cybersecurity issues that are not well addressed by commonly employed security solutions. In this work, we propose structural taxonomy and novel categorization of research that contribute to the organization and disambiguation of insider threat incidents and the defense solutions used against them. The objective of our categorization is to systematize knowledge in insider threat research while using an existing grounded theory method for rigorous literature review. The proposed categorization depicts the workflow among particular categories that include incidents and datasets, analysis of incidents, simulations, and defense solutions. Special attention is paid to the definitions and taxonomies of the insider threat; we present a structural taxonomy of insider threat incidents that is based on existing taxonomies and the 5W1H questions of the information gathering problem. Our survey will enhance researchers’ efforts in the domain of insider threat because it provides (1) a novel structural taxonomy that contributes to orthogonal classification of incidents and defining the scope of defense solutions employed against them, (2) an overview on publicly available datasets that can be used to test new detection solutions against other works, (3) references of existing case studies and frameworks modeling insiders’ behaviors for the purpose of reviewing defense solutions or extending their coverage, and (4) a discussion of existing trends and further research directions that can be used for reasoning in the insider threat domain.</jats:p>

<jats:p>The common perception in both academic literature and industry today is that virtual machines offer better security, whereas containers offer better performance. However, a detailed review of the history of these technologies and the current threats they face reveals a different story. This survey covers key developments in the evolution of virtual machines and containers from the 1950s to today, with an emphasis on countering modern misperceptions with accurate historical details and providing a solid foundation for ongoing research into the future of secure isolation for multitenant infrastructures, such as cloud and container deployments.</jats:p>

<jats:p> In the absence of any global positioning infrastructure for indoor environments, research on supporting human indoor localization and navigation trails decades behind research on outdoor localization and navigation. The major barrier to broader progress has been the dependency of indoor positioning on environment-specific infrastructure and resulting tailored technical solutions. Combined with the fragmentation and compartmentalization of indoor environments, this poses significant challenges to widespread adoption of indoor location-based services. This article puts aside all approaches of infrastructure-based support for human indoor localization and navigation and instead reviews technical concepts that are <jats:italic>independent</jats:italic> of sensors embedded in the environment. The reviewed concepts rely on a mobile computing platform with sensing capability and a human interaction interface (“smartphone”). This platform may or may not carry a stored map of the environment, but does not require <jats:italic>in situ</jats:italic> internet access. In this regard, the presented approaches are more challenging than any localization and navigation solutions specific to a particular, infrastructure-equipped indoor space, since they are not adapted to local context, and they may lack some of the accuracy achievable with those tailored solutions. However, only these approaches have the potential to be universally applicable. </jats:p>

<jats:p>The Information-Centric Network (ICN) paradigm is a future Internet approach aiming to tackle the Internet architectural problems and inefficiencies, by swapping the main entity of the network architecture from hosts to content items. In ICN, content names play a central role: Each content gets a unique name at the network layer, and this name is used for routing the content over the network. This paradigm change potentially enables a future Internet with better performance, reliability, scalability, and suitability for wireless and mobile communication. It also provides new intrinsic means to deal with some popular attacks on the Internet architecture, such as denial of service. However, this new paradigm also represents new challenges related to security that need to be addressed, to ensure its capability to support current and future Internet requirements. This article surveys and summarizes ongoing research concerning security aspects of ICNs, discussing vulnerabilities, attacks, and proposed solutions to mitigate them. We also discuss open challenges and propose future directions regarding research in ICN security.</jats:p>

<jats:p>Data prefetching, i.e., the act of predicting an application’s future memory accesses and fetching those that are not in the on-chip caches, is a well-known and widely used approach to hide the long latency of memory accesses. The fruitfulness of data prefetching is evident to both industry and academy: Nowadays, almost every high-performance processor incorporates a few data prefetchers for capturing various access patterns of applications; besides, there is a myriad of proposals for data prefetching in the research literature, where each proposal enhances the efficiency of prefetching in a specific way.</jats:p> <jats:p>In this survey, we evaluate the effectiveness of data prefetching in the context of server applications and shed light on its design trade-offs. To do so, we choose a target architecture based on a contemporary server processor and stack various state-of-the-art data prefetchers on top of it. We analyze the prefetchers in terms of their ability to predict memory accesses and enhance overall system performance, as well as their imposed overheads. Finally, by comparing the state-of-the-art prefetchers with impractical ideal prefetchers, we motivate further work on improving data prefetching techniques.</jats:p>

<jats:p>Multisensory experiences have been increasingly applied in Human-Computer Interaction (HCI). In recent years, it is commonplace to notice the development of haptic, olfactory, and even gustatory displays to create more immersive experiences. Companies are proposing new additions to the multisensory world and are unveiling new products that promise to offer amazing experiences exploiting mulsemedia—multiple sensorial media—where users can perceive odors, tastes, and the sensation of wind blowing against their face. Whilst researchers, practitioners and users alike are faced with a wide range of such new devices, relatively little work has been undertaken to summarize efforts and initiatives in this area. The current article addresses this shortcoming in two ways: first, by presenting a survey of devices targeting senses beyond that of sight and hearing and, second, by describing an approach to guide newcomers and experienced practitioners alike to build their own mulsemedia environment, both in a desktop setting and in an immersive 360° environment.</jats:p>

<jats:p>Integration of embedded multimedia devices with powerful computing platforms, e.g., machine learning platforms, helps to build smart cities and transforms the concept of Internet of Things into Internet of Multimedia Things (IoMT). To provide different services to the residents of smart cities, the IoMT technology generates big multimedia data. The management of big multimedia data is a challenging task for IoMT technology. Without proper management, it is hard to maintain consistency, reusability, and reconcilability of generated big multimedia data in smart cities. Various machine learning techniques can be used for automatic classification of raw multimedia data and to allow machines to learn features and perform specific tasks. In this survey, we focus on various machine learning platforms that can be used to process and manage big multimedia data generated by different applications in smart cities. We also highlight various limitations and research challenges that need to be considered when processing big multimedia data in real-time.</jats:p>

<jats:p>Group key agreement (shorten as GKA) protocol enables a group of users to negotiate a one-time session key and protect the thereafter group-oriented communication with this session key across an unreliable network. The number of communication rounds is one of the main concern for practical applications where the cardinality of group participants involved is considerable. It is critical to have fixed constant rounds in GKA protocols to secure these applications. In light of overwhelming variety and multitude of constant-round GKA protocols, this article surveys these protocols from a series of perspectives to supply better comprehension for researchers and scholars. Concretely, this article captures the state of the art of constant-round GKA protocols by analyzing the design rationale, examining the framework and security model, and evaluating all discussed protocols in terms of efficiency and security properties. In addition, this article discusses the extension of constant-round GKA protocols including dynamic membership updating, password-based, affiliation-hiding, and fault-tolerance. In conclusion, this article also points out a number of interesting future directions.</jats:p>