Search

Domain generalization models learn to generalize to previously unseen domains, but suffer from prediction uncertainty and domain shift. In this paper, we address both problems. We introduce a probabilistic meta-learning model for domain generalization, in which classifier parameters shared across domains are modeled as distributions. This enables better handling of prediction uncertainty on unseen domains. To deal with domain shift, we learn domain-invariant representations by the proposed principle of meta variational information bottleneck, we call MetaVIB. MetaVIB is derived from novel variational bounds of mutual information, by leveraging the meta-learning setting of domain generalization. Through episodic training, MetaVIB learns to gradually narrow domain gaps to establish domain-invariant representations, while simultaneously maximizing prediction accuracy. We conduct experiments on three benchmarks for cross-domain visual recognition. Comprehensive ablation studies validate the benefits of MetaVIB for domain generalization. The comparison results demonstrate our method outperforms previous approaches consistently.

Küçükkoç, İbrahim (Balikesir Author), Typically, the total number of required workstations are minimised for a given cycle time (this problem is referred to as type-1), or cycle time is minimised for a given number of workstations (this problem is referred to as type-2) in traditional balancing of assembly lines. However, variation in workload distributions of workstations is an important indicator of the quality of the obtained line balance. This needs to be taken into account to improve the reliability of an assembly line against unforeseeable circumstances, such as breakdowns or other failures. For this aim, a new problem, called lexicographic bottleneck mixed-model assembly line balancing problem (LB-MALBP), is presented and formalised. The lexicographic bottleneck objective, which was recently proposed for the simple single-model assembly line system in the literature, is considered for a mixed-model assembly line system. The mathematical model of the LB-MALBP is developed for the first time in the literature and coded in GAMS solver, and optimal solutions are presented for some small scale test problems available in the literature. As it is not possible to get optimal solutions for the large-scale instances, an artificial bee colony algorithm is also implemented for the solution of the LB-MALBP. The solution procedures of the algorithm are explored illustratively. The performance of the algorithm is also assessed using derived well-known test problems in this domain and promising results are observed in reasonable CPU times.

4OR (2021). doi:10.1007/s10288-021-00499-6, Published by Springer, Berlin ; Heidelberg

In the past decade, advances in the speed of commodity CPUs have far out-paced advances in memory latency. Main-memory access is therefore increasingly a performance bottleneck for many computer applications, including database systems. In this article, we use a simple scan test to show the severe impact of this bottleneck. The insights gained are translated into guidelines for database architecture, in terms of both data structures and algorithms. We discuss how vertically fragmented data structures optimize cache performance on sequential data access. We then focus on equi-join, typically a random-access operation, and introduce radix algorithms for partitioned hash-join. The performance of these algorithms is quantified using a detailed analytical model that incorporates memory access cost. Experiments that validate this model were performed on the Monet database system. We obtained exact statistics on events such as TLB misses and L1 and L2 cache misses by using hardware performance counters found in modern CPUs. Using our cost model, we show how the carefully tuned memory access pattern of our radix algorithms makes them perform well, which is confirmed by experimental results.

Acoustic indices derived from environmental soundscape recordings are being used to monitor ecosystem health and vocal animal biodiversity. Soundscape data can quickly become very expensive and difficult to manage, so data compression or temporal down‐sampling are sometimes employed to reduce data storage and transmission costs. These parameters vary widely between experiments, with the consequences of this variation remaining mostly unknown.We analyse field recordings from North‐Eastern Borneo across a gradient of historical land use. We quantify the impact of experimental parameters (MP3 compression, recording length and temporal subsetting) on soundscape descriptors (Analytical Indices and a convolutional neural net derived AudioSet Fingerprint). Both descriptor types were tested for their robustness to parameter alteration and their usability in a soundscape classification task.We find that compression and recording length both drive considerable variation in calculated index values. However, we find that the effects of this variation and temporal subsetting on the performance of classification models is minor: performance is much more strongly determined by acoustic index choice, with Audioset fingerprinting offering substantially greater (12%–16%) levels of classifier accuracy, precision and recall.We advise using the AudioSet Fingerprint in soundscape analysis, finding superior and consistent performance even on small pools of data. If data storage is a bottleneck to a study, we recommend Variable Bit Rate encoded compression (quality = 0) to reduce file size to 23% file size without affecting most Analytical Index values. The AudioSet Fingerprint can be compressed further to a Constant Bit Rate encoding of 64 kb/s (8% file size) without any detectable effect. These recommendations allow the efficient use of restricted data storage whilst permitting comparability of results between different studies., Soundscapes were recorded from different forest structures in Malaysian Borneo. Data collection variation was simulated, and all data groups were analyzed via usual acoustic indices and a CNN‐derived AudioSet Fingerprint. The effect of variation in data collection was compared between the two types of soundscape descriptor, finding the AudioSet Fingerprint to be a stronger and more robust descriptor of soundscapes.

Analysis of cardiac ultrasound images is commonly performed in routine clinical practice for quantification of cardiac function. Its increasing automation frequently employs deep learning networks that are trained to predict disease or detect image features. However, such models are extremely data-hungry and training requires labelling of many thousands of images by experienced clinicians. Here we propose the use of contrastive learning to mitigate the labelling bottleneck. We train view classification models for imbalanced cardiac ultrasound datasets and show improved performance for views/classes for which minimal labelled data is available. Compared to a naive baseline model, we achieve an improvement in F1 score of up to 26% in those views while maintaining state-of-the-art performance for the views with sufficiently many labelled training observations., Accepted in ASMUS workshop of MICCAI 2021

This paper presents a computer tool for collaborative narration creation based on graph rules hierarchy that was developed and tested with collaboration of students of Jagiellonian University in Poland. The tool is made of two modules: internal representation of game world graph model and animation system: The former describes the narration points and story arcs and the latter uses basic animations and blends them together to create animated sequences. The opportunity to develop the system as a group effort allowed to identify and address workflow bottleneck and proved that using layered graph representation of the functional elements of the story, characters, locations, and objects is viable method to design plots and quest for video games. Dynamically changing graph can be visualized at any stage and helps to perceive the story as animated sequences. This on the other hand if used for video games story design enables to observe character actions, possible paths the player can take and quickly iterate on the story components. The contribution of this paper is to propose an innovative tool which allow the designers and animators to work together within one platform.

With increasing data and model complexities, the time required to train neural networks has become prohibitively large. To address the exponential rise in training time, users are turning to data parallel neural networks (DPNN) and large-scale distributed resources on computer clusters. Current DPNN approaches implement the network parameter updates by synchronizing and averaging gradients across all processes with blocking communication operations after each forward-backward pass. This synchronization is the central algorithmic bottleneck. We introduce the Distributed Asynchronous and Selective Optimization (DASO) method, which leverages multi-GPU compute node architectures to accelerate network training while maintaining accuracy. DASO uses a hierarchical and asynchronous communication scheme comprised of node-local and global networks while adjusting the global synchronization rate during the learning process. We show that DASO yields a reduction in training time of up to 34% on classical and state-of-the-art networks, as compared to current optimized data parallel training methods.

The file attached to this record is the author's final peer reviewed version. The Publisher's final version can be found by following the DOI link. Introduction Metabolomics is the approach of choice to guide the understanding of biological systems and its molecular intricacies, but compound identification is yet a bottleneck to be overcome. Objective To assay the use of NMRfilter for confidence compound identification based on chemical shift predictions for different datasets. Results We found comparable results using the lead tool COLMAR and NMRfilter. Then, we successfully assayed the use of HMBC to add confidence to the identified compounds. Conclusions NMRfilter is currently under development to become a stand-alone interactive software for high-confidence NMR compound identification and this communication gathers part of its application capabilities.

Monte Carlo simulation is a common technique to estimate dependability metrics for fault trees. A bottleneck in this technique is the number of samples needed, especially when the interesting events are rare and occur with low probability. Rare Event Simulation (Open image in new window) reduces the number of samples when analysing rare events. Importance splitting is a Open image in new window method that spawns more simulation runs from promising system states. How promising a state is, is indicated by an importance function, which concentrates the information that makes this method efficient. Importance functions are given by domain and Open image in new window experts. This hinders re-utilisation and involves decisions entailing potential human error. Focusing in (general) fault trees, in this paper we automatically derive importance functions based on the tree structure. For this we exploit a common fault tree concept, namely cut sets: the more elements from a cut set have failed, the higher the importance. We show that the cut-set-derived importance function is an easy-to-implement and simple concept, that can nonetheless compete against another (more involved) automatic importance function for Open image in new window.

This paper describes how a time-based planning system, which supports resource constraints, may be extended such that a resource constraint interval does not have to refer to the start- or end-time of the underlying activity but to any linear combination thereof, such as the middle. This way, an activity with multiple resource constraints referring to different time intervals no longer has to be split into sub-activities, which may simplify the planning model and the algorithm. To be able to describe the necessary transformations, we introduce the concept of PolygonStacks and describe the operations which a typical planning engine requires to intersect the sets of consistent timeline entries of all constraints defined on an activity. We then introduce Sliders and Offsets, which allow specifying the constraint intervals in a more generic way as supported in current planning models. Based on this preparation, we can derive two lemmas, which provide the conversions required by Sliders and Offsets. We continue with several conversion examples and point out how to solve the issues which will occur during implementation. A short sketch of the complexity of our current implementation demonstrates that further work on performance should be considered, even though in practice we observe that the bottleneck of calculation remains within profile calculation rather than PolygonStack operations.

Zero-knowledge protocols enable the truth of a mathematical statement to be certified by a verifier without revealing any other information. Such protocols are a cornerstone of modern cryptography and recently are becoming more and more practical. However, a major bottleneck in deployment is the efficiency of the prover and, in particular, the space-efficiency of the protocol.

Biomolecular computing is the field of engineering where computation, storage, communication, and coding are obtained by exploiting interactions between biomolecules, especially DNA, RNA, and enzymes. They are a promising solution in a long-term vision, bringing huge parallelism and negligible power consumption. Despite significant efforts in taking advantage of the massive computational power of biomolecules, many issues are still open along the way for considering biomolecular circuits as an alternative or a complement to competing with complementary metal–oxide–semiconductor (CMOS) architectures. According to the Von Neumann architecture, computing systems are composed of a central processing unit, a storage unit, and input and output (I/O). I/O operations are crucial to drive and read the computing core and to interface it to other devices. In emerging technologies, the complexity overhead and the bottleneck of I/O systems are usually limiting factors. While computing units and memories based on biomolecular systems have been successfully presented in literature, the published I/O operations are still based on laboratory equipment without a real development of integrated I/O. Biosensors are suitable devices for transducing biomolecular interactions by converting them into electrical signals. In this work, we explore the latest advancements in biomolecular computing, as well as in biosensors, with focus on technology suitable to provide the required and still missing I/O devices. Therefore, our goal is to picture out the present and future perspectives about DNA, RNA, and enzymatic-based computing according to the progression in its I/O technologies, and to understand how the field of biosensors contributes to the research beyond CMOS.

This paper focuses on challenging applications that can be expressed as an iterative pipeline of multiple 3d stencil stages and explores their optimization space on GPUs. For this study, we selected a representative example from the field of digital signal processing, the Anisotropic Nonlinear Diffusion algorithm. An open issue to these applications is to determine the optimal fission/fusion level of the involved stages and whether that combination benefits from data tiling. This implies exploring a large space of all the possible fission/fusion combinations with and without tiling, thus making the process non-trivial. This study provides insights to reduce the optimization tuning space and programming effort of iterative multiple 3d stencils. Our results demonstrate that all combinations that fuse the bottleneck stencil with high halos update cost ( $$>25\%$$ , this percentage can be measured or estimated experimentally for each single stencil) and high registers and shared memory accesses must not be considered in the exploration process. The optimal fission/fusion combination is up to 1.65 $$\times $$ faster than the case in which we fully decompose our stencil without tiling and 5.3 $$\times $$ faster with respect to the fully fused version on the NVIDIA GPUs.

beta-defensins are important components of the vertebrate innate immune system responsible for encoding a variety of anti-microbial peptides. Pathogen-mediated selection is thought to act on immune genes and potentially maintain allelic variation in the face of genetic drift. The Seychelles warbler, Acrocephalus sechellensis, is an endemic passerine that underwent a recent bottleneck in its last remaining population, resulting in a considerable reduction in genome-wide variation. We genotyped avian beta-defensin (AvBD) genes in contemporary (2000-2008) and museum samples (1876-1940) of the Seychelles warbler to investigate whether immunogenetic variation was lost through this bottleneck, and examined AvBD variation across four other Acrocephalus species with varying demographic histories. No variation was detected at four of the six AvBD loci screened in the post-bottleneck population of Seychelles warbler, but two silent nucleotide polymorphisms were identified at AvBD8 and one potentially functional amino-acid variation was observed at AvBD11. Variation in the Seychelles warbler was significantly lower than in the mainland migratory congeneric species investigated, but it similar to that found in other bottlenecked species. In addition, screening AvBD7 in 15 museum specimens of Seychelles warblers sampled prior to the bottleneck (1877-1905) revealed that this locus possessed two alleles previously, compared to the single allele in the contemporary population. Overall, the results show that little AvBD variation remains in the Seychelles warbler, probably as a result of having low AvBD diversity historically rather than the loss of variation due to drift associated with past demographic history. Given the limited pathogen fauna, this lack of variation at the AvBD loci may currently not pose a problem for this isolate population of Seychelles warblers, but it may be detrimental to the species' long-term survival if new pathogens reach the population in the future.

For the synthesis of correct-by-construction control policies from temporal logic specifications the scalability of the synthesis algorithms is often a bottleneck. In this paper, we parallelize synthesis from specifications in the GR(1) fragment of linear temporal logic by introducing a hierarchical procedure that allows decoupling of the fixpoint computations. The state space is partitioned into equicontrollable sets using solutions to parametrized games that arise from decomposing the original GR(1) game into smaller reachability-persistence games. Following the partitioning, another synthesis problem is formulated for composing the strategies from the decomposed reachability games. The formulation guarantees that composing the synthesized controllers ensures satisfaction of the given GR(1) property. Experiments with robot planning problems demonstrate good performance of the approach.

Recovering brain connectivity from tract tracing data is an important computational problem in the neurosciences. Mesoscopic connectome reconstruction was previously formulated as a structured matrix regression problem (Harris et al. in Neural Information Processing Systems, 2016), but existing techniques do not scale to the whole-brain setting. The corresponding matrix equation is challenging to solve due to large scale, ill-conditioning, and a general form that lacks a convergent splitting. We propose a greedy low-rank algorithm for the connectome reconstruction problem in very high dimensions. The algorithm approximates the solution by a sequence of rank-one updates which exploit the sparse and positive definite problem structure. This algorithm was described previously (Kressner and Sirković in Numer Lin Alg Appl 22(3):564–583, 2015) but never implemented for this connectome problem, leading to a number of challenges. We have had to design judicious stopping criteria and employ efficient solvers for the three main sub-problems of the algorithm, including an efficient GPU implementation that alleviates the main bottleneck for large datasets. The performance of the method is evaluated on three examples: an artificial “toy” dataset and two whole-cortex instances using data from the Allen Mouse Brain Connectivity Atlas. We find that the method is significantly faster than previous methods and that moderate ranks offer a good approximation. This speedup allows for the estimation of increasingly large-scale connectomes across taxa as these data become available from tracing experiments. The data and code are available online. ispartof: Journal Of Mathematical Neuroscience vol:9 issue:1 ispartof: location:Germany status: Published online

The parallel computation capabilities of modern GPU (Graphics Processing Unit) processors have attracted increasing attention from researchers and engineers who have been conducting high computational throughput studies. However, current single GPU based engineering solutions are often struggle to fulfill their real-time requirements. Thus, the multi-GPU-based approach has become a popular and cost-effective choice for tackling the demands. In those cases, the computational load balancing over multiple GPU “nodes” is often the key and bottleneck that affect the quality and performance of the runtime system. The existing load balancing approaches are mainly based on the assumption that all GPU nodes in the same computer framework are of equal computational performance, which are often not the case due to cluster design and other legacy issues. This paper presents a novel dynamic load balancing (DLB) model for rapid data division and allocation on heterogeneous GPU nodes based on an innovative fuzzy neural network (FNN). In this research, a 5-state parameter feedback mechanism defining the overall cluster and node performances is proposed. The corresponding FNN-based DLB model will be capable of monitoring and predicting individual node performance under different workload scenarios. A real-time adaptive scheduler has been devised to reorganize the data inputs to each node when necessary to maintain their runtime computational performances. The devised model has been implemented on two dimensional (2D) discrete wavelet transform (DWT) tasks for evaluation. Experiment results show that this DLB model has enabled a high computational throughput while ensuring real-time and precision requirements from complex computational tasks.

The high dispersal ability of seabirds and theabsence of geographical barriers has led to high gene flowand reduced population differentiation. Nevertheless, somespecies with philopatric behavior have restricted gene flowamong colonies, revealing a strong population structure.Gulls show widespread colonial behavior, and are longlivedspecies, which make them a good model for understandingevolutionary processes in seabirds. Previousgenetic studies on the Kelp Gull (Larus dominicanus) haverevealed low genetic variability in mitochondrial markersbut relatively high genetic variation in a nuclear marker.These observations can be explained by the occurrence of aselective sweep on mtDNA, population genetic bottlenecksor a recent origin of the species. We used microsatellitedata to further investigate these hypotheses, mainly bytesting for bottleneck events. Low genetic variability(Ho = 0.276?0.570) was detected in Kelp Gulls. However,population genetic structure was observed among regions(Chile, Argentina and Brazil), and between continents(South America and Antarctica). The population of theKelp Gull in South America may have differentiated due toisolation by distance (r = 0.7273, p = 0.0013), whereasthe population in the Antarctic seems to be isolated by nonphysicalbarriers. Bottleneck events were detected in 6 outof 14 colonies studied. These colonies are at the limits ofthe distribution of the Kelp Gull, and thus experience harshsurvival conditions. We believe that the Kelp Gull has acomplex history in the southern hemisphere, with a recentorigin, followed by bottlenecks and then population expansion. Thus, the genetic diversity found in Kelp Gull issimilar to that observed for other species of Laridae. Fil: de Almeida Santos, Fernanda. Universidade de Sao Paulo; Brasil Fil: Stenghel Morgante, João. Universidade de Sao Paulo; Brasil Fil: Frere, Esteban. Universidad Nacional de la Patagonia Austral; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina Fil: Millones, Ana. Universidad Nacional de la Patagonia Austral; Argentina Fil: Sander, Martin. Universidad de Vale do Rio dos Sinos; Brasil Fil: de Abreu Vianna, Juliana. Pontificia Universidad Católica de Chile; Chile Fil: Pires de Mendonça Dantas, Gisele. Universidade de Sao Paulo; Brasil. Pontificia Universidade Catolica de Minas Gerais.; Brasil

Migration monitoring may allow us to detect population trends over large geographic areas because the pattern of change in migrant counts may be expected to follow the pattern of change in population size. We analysed recent regional European population trends of migratory soaring birds from rates of change in migration counts over the Strait of Gibraltar (Spain) during the years (1999–2013). An additional bottleneck (Organbidexka, France) within the same migratory route and period was also considered. We estimated count trends by fitting a log-generalized linear model to the time series of each species counts. The counts in Organbidexka were used to test the consistency in the observed trends over the Strait of Gibraltar. Migration counts of White and Black Storks, Black Kites, Short-toed and Booted Eagles as well as Egyptian Vultures showed a linear increase over the Strait of Gibraltar throughout the 15-year period. In contrast, Honey Buzzard numbers remained stable. Trends were highly consistent with those recorded in Organbidexka. We suggest that the larger slopes for the trends in Organbidexka when compared with the Strait reflect an increasing tendency in these species to overwinter in southern Europe. A combination of complementary data sets collected at different bottleneck sites within the European–African flyway system may become a fundamental tool for the investigation of migratory patterns and population trends and changes of soaring migrant birds all over Europe.https://doi.org/10.1007/s10531-016-1047-4, The data counts from the Strait of Gibraltar analyzed in the study were collected in field monitoring campaigns 1999/2012 funded by grants of the Consejeria de Medio Ambiente of the Junta de Andalucía (Spain)

Deployment and maintenance of current data center networks is costly and prone to errors. In order to avoid manual configuration, many of them require centralized administrators which constitute a clear bottleneck, while distributed approaches do not guarantee sufficient flexibility or robustness. This paper describes and evaluates GA3 (Generalized Automatic Address Assignment), a discovery protocol that assigns multiple unique labels to all the switches in a hierarchical network, without any modification of hosts or the standard Ethernet frames. Labeling is distributed and uses probes. These labels can be leveraged for shortest path routing without tables, as in the case of the Torii protocol, but GA3 also allows other label-based routing protocols (such as PortLand or ALIAS). Additionally, GA3 can detect miswirings in the network. Furthermore, control traffic is only necessary upon network deployment rather than periodically. Simulation results showed a reduced convergence time of less than 2 s and 100 kB/s of bandwidth (to send the GA3 frames) in the worst case for popular data center topologies, which outperforms other similar protocols., Comunidad de Madrid

Microalgae have been proposed as an option for wastewater treatment since the 1960s, but still, this technology has not been expanded to an industrial scale. In this paper, the major factors limiting the performance of these systems are analysed. The composition of the wastewater is highly relevant, and especially the presence of pollutants such as heavy metals and emerging compounds. Biological and engineering aspects are also critical and have to be improved to at least approximate the performance of conventional systems, not just in terms of capacity and efficiency but also in terms of robustness. Finally, the harvesting of the biomass and its processing into valuable products pose a challenge; yet at the same time, an opportunity exists to increase economic profitability. Land requirement is a major bottleneck that can be ameliorated by improving the system’s photosynthetic efficiency. Land requirement has a significant impact on the economic balance, but the profits from the biomass produced can enhance these systems’ reliability, especially in small cities.

12 páginas, 4 figuras, 4 tablas.-- et al., Monitoring the loss of genetic diversity in wild populations after a bottleneck event is a priority in conservation and management plans. Here, we used diverse molecular markers to search for signatures of demographic bottlenecks in two wolf populations; an isolated population from the Iberian Peninsula and a non-isolated population from European Russia. Autosomal, mtDNA and Y-chromosomal diversity and the effective population size (Ne) were significantly lower in the Iberian population. Neutrality tests using mtDNA sequences, such as R2, Fu and Li’s F*, Tajima’s D and Fu’s Fs, were positively significant in the Iberian population, suggesting a population decline, but were not significant for the Russian population, likely due to its larger effective population size. However, three tests using autosomal data confirmed the occurrence of the genetic bottleneck in both populations. The M-ratio test was the only one providing significant results for both populations. Given the lack of consistency among the different tests, we recommend using multiple approaches to investigate possible past bottlenecks. The small effective population size (about 50) in the Iberian Peninsula compared to the presumed extant population size could indicate that the bottleneck was more powerful than initially suspected or an overestimation of the current population. The risks associated with small effective population sizes suggest that the genetic change in this population should be closely monitored in the future. On the other hand, the relatively small effective population size for Russian wolves (a few hundred individuals) could indicate some fragmentation, contrary to what is commonly assumed., Carles Vilà work was supported by the ‘‘Programa para la Captación del Conocimiento para Andalucía’’ (Andalusian Government, Spain). Financial support was provided by the ‘‘Servei Veterinari de Genètica Molecular’’ (SVGM).

Transport modelling and in particular transport assignment is a well-known bottleneck in computation cost and time for urban system models. The use of Transport Analysis Zones (TAZ) implies a trade-off between computation time and accuracy: practical computational constraints can lead to concessions to zone size with severe repercussions for the quality of the transport representation in urban models. This paper investigates how a recently developed geographical topology called adaptive zoning can be used to obtain more favorable trade-offs between computational cost and accuracy than traditional TAZ. Adaptive zoning was developed specifically for representing spatial interactions; it makes use of a nested zone hierarchy to adapt the model resolution as a function of both the origin and destination location. In this paper the adaptive zoning method is tied to an approach to trip assignment that uses high spatial accuracy (small zones) at one end of the route and low spatial accuracy (large zones) at the other end of the route. Opportunistic use of either the first or second half of such routes with asymmetric accuracy profiles leads to a method of transport assignment that is more accurate than traditional TAZ based assignment at reduced computational cost. The method is tested and demonstrated on the well-known Chicago Regional test problem. Compared with an assignment using traditional zoning, an adaptive-zoning-based assignment that uses the same computation time reduces the bias in travel time by a factor 16 and link level traffic volume RMSE by a factor 6.4.

© The Editor(s) (if applicable) and the Author(s) 2015. While many innovations in Technology Enhanced Learning (TEL) have emerged over the last two decades, the uptake of these innovations has not always been very successful, particularly in schools. The transition from proof of concept to integration into learning activities has been recognized as a bottleneck for quite some time. This major problem, which is affecting many TEL stakeholders, is the focus of the 4-year iTEC project that is developing a comprehensive approach involving 15 ministries of education and is organizing a large scale validator with more than a thousand classrooms. This chapter reports on how the information provision on events of interest in learning as well as on persons that can contribute to learning activities, supports novel scenarios and is key for the introduction of open education in the K12 education. ispartof: Re-engineering the Uptake of ICT in Schools pages:115-139 ispartof: pages:115-139 status: published

The selection of points-of-interest in leakage traces is a frequently neglected problem in the side-channel literature. However, it can become the bottleneck of practical adversaries/evaluators as the size of the measurement traces increases, especially in the challenging context of masked implementations, where only a combination of multiple shares reveals information in higher-order statistical moments. In this paper, we describe new black box tools for efficiently dealing with this problem. The proposed techniques exploit projection pursuits and specialized local search algorithms, work with minimum memory requirements and practical time complexity. We validate them with two case-studies of unprotected and first-order masked implementations in an 8-bit device, the latter one being hard to analyze with previously known methods.

This paper presents an adaptive resource control mechanism for multi-hop ad-hoc network systems, which avoids bottleneck problems caused by the node-fairness property of IEEE 802.11. In our proposal, the feedback information from the downstream bottleneck, derived from Request-To-Send (RTS) and Clear-To-Send (CTS) messages is utilized to control the Transmission Opportunity (TXOP) limit of the upstream nodes for traffic balancing. The proposed mechanism is modelled control-theoretically using the 20-sim control system modelling tool, which has the advantage that results can be obtained in a fast and efficient way. Compared to systems without resource control, a higher throughput and lower delay can be achieved under a variety of traffic load conditions as well as in dynamic network environments.

Two-hop ad-hoc networks, in which some nodes forward traffic for multiple sources, with which they also compete for channel access suffer from large queues building up in bottleneck nodes. This problem can often be alleviated by using IEEE 802.11e to give preferential treatment to bottleneck nodes. Previous results have shown that differentiation parameters can be used to allocate capacity in a more efficient way in the two-hop scenario. However, the overall throughput of the bottleneck may differ considerably, depending on the differentiation method used. By applying a very fast and accurate analysis method, based on steady-state analysis of an QBD-type infinite Markov chain, we find the maximum throughput that is possible per differentiation parameter. All possible parameter settings are explored with respect to the maximum throughput conditioned on a maximum buffer occupancy. This design space exploration cannot be done with network simulators like NS2 or Opnet, as each simulation run simply takes to long. The results, which have been validated by detailed simulations, show that by differentiating TXOP it is possible to achieve a throughput that is about 50% larger than when differentiating AIFS and $\text{CW}_{\text{min}}$.

The Linked Data cloud has grown to become the largest knowledge base ever constructed. Its size is now turning into a major bottleneck for many applications. In order to facilitate access to this structured information, this paper proposes an automatic sampling method targeted at maximizing answer coverage for applications using SPARQL querying. The approach presented in this paper is novel: no similar RDF sampling approach exist. Additionally, the concept of creating a sample aimed at maximizing SPARQL answer coverage, is unique. We empirically show that the relevance of triples for sampling (a semantic notion) is influenced by the topology of the graph (purely structural), and can be determined without prior knowledge of the queries. Experiments show a significantly higher recall of topology based sampling methods over ran- dom and naive baseline approaches (e.g. up to 90% for Open-BioMed at a sample size of 6%).

Wireless ad-hoc networks are based on shared medium technology where the nodes arrange access to the medium in a distributed way independent of their current traffic demand. This has the inherent drawback that a node that serves as a relay node for transmissions of multiple neighboring nodes is prone to become a performance "bottleneck". In the present paper such a bottleneck node is modeled via an idealized fluid-flow queueing model in which the complex packet-level behavior (MAC) is represented by a small set of parameters. We extensively validate the model by ad-hoc network simulations that include all the details of the widely used IEEE 802.11 MAC-protocol. Further we show that the overall flow transfer time of a multi-hop flow, which consists of the sum of the delays at the individual nodes, improves by granting a larger share of the medium capacity to the bottleneck node. Such alternative resource sharing strategies can be enforced in real systems by deploying the recently standardized IEEE 802.HE MAC-protocol. We propose a mapping between the parameter settings of IEEE 802. HE and the fluid-flow model, and validate the fluid-flow model and the parameter mapping with detailed system simulations. © Springer-Verlag Berlin Heidelberg 2007.