Search

Learning the structure of a Bayesian Network (BN) with score-based solutions involves exploring the search space of possible graphs and moving towards the graph that maximises a given objective function. Some algorithms offer exact solutions that guarantee to return the graph with the highest objective score, while others offer approximate solutions in exchange for reduced computational complexity. This paper describes an approximate BN structure learning algorithm, which we call Model Averaging Hill-Climbing (MAHC), that combines two novel strategies with hill-climbing search. The algorithm starts by pruning the search space of graphs, where the pruning strategy can be viewed as an aggressive version of the pruning strategies that are typically applied to combinatorial optimisation structure learning problems. It then performs model averaging in the hill-climbing search process and moves to the neighbouring graph that maximises the objective function, on average, for that neighbouring graph and over all its valid neighbouring graphs. Comparisons with other algorithms spanning different classes of learning suggest that the combination of aggressive pruning with model averaging is both effective and efficient, particularly in the presence of data noise.

Learning from data that contain missing values represents a common phenomenon in many domains. Relatively few Bayesian Network structure learning algorithms account for missing data, and those that do tend to rely on standard approaches that assume missing data are missing at random, such as the Expectation-Maximisation algorithm. Because missing data are often systematic, there is a need for more pragmatic methods that can effectively deal with data sets containing missing values not missing at random. The absence of approaches that deal with systematic missing data impedes the application of BN structure learning methods to real-world problems where missingness are not random. This paper describes three variants of greedy search structure learning that utilise pairwise deletion and inverse probability weighting to maximally leverage the observed data and to limit potential bias caused by missing values. The first two of the variants can be viewed as sub-versions of the third and best performing variant, but are important in their own in illustrating the successive improvements in learning accuracy. The empirical investigations show that the proposed approach outperforms the commonly used and state-of-the-art Structural EM algorithm, both in terms of learning accuracy and efficiency, as well as both when data are missing at random and not at random.

In this work, we show the feasibility of extreme ultraviolet (EUV) patterning on an HF-treated silicon (100) surface in the absence of a photoresist. EUV lithography is the leading lithography technique in semiconductor manufacturing due to its high resolution and throughput, but future progress in resolution can be hampered because of the inherent limitations of the resists. We show that EUV photons can induce surface reactions on a partially hydrogen-terminated silicon surface and assist the growth of an oxide layer, which serves as an etch mask. This mechanism is different from the hydrogen desorption in scanning tunneling microscopy-based lithography. We achieve silicon dioxide/silicon gratings with 75-nanometer half-pitch and 31-nanometer height, demonstrating the efficacy of the method and the feasibility of patterning with EUV lithography without the use of a photoresist. Further development of the resistless EUV lithography method can be a viable approach to nanometer-scale lithography by overcoming the inherent resolution and roughness limitations of photoresist materials., Science Advances, 9 (16), ISSN:2375-2548

The progress of miniaturisation in integrated electronics has led to atomic and nanometre-sized dopant devices in silicon. Such structures can be fabricated routinely by hydrogen resist lithography, using various dopants such as phosphorous and arsenic. However, the ability to non-destructively obtain atomic-species-specific images of the final structure, which would be an indispensable tool for building more complex nano-scale devices, such as quantum co-processors, remains an unresolved challenge. Here we exploit X-ray fluorescence to create an element-specific image of As dopants in silicon, with dopant densities in absolute units and a resolution limited by the beam focal size (here $\sim1~\mu$m), without affecting the device's low temperature electronic properties. The As densities provided by the X-ray data are compared to those derived from Hall effect measurements as well as the standard non-repeatable, scanning tunnelling microscopy and secondary ion mass spectroscopy, techniques. Before and after the X-ray experiments, we also measured the magneto-conductance, dominated by weak localisation, a quantum interference effect extremely sensitive to sample dimensions and disorder. Notwithstanding the $1.5\times10^{10}$ Sv ($1.5\times10^{16}$ Rad/cm$^{-2}$) exposure of the device to X-rays, all transport data were unchanged to within experimental errors, corresponding to upper bounds of 0.2 Angstroms for the radiation-induced motion of the typical As atom and 3$\%$ for the loss of activated, carrier-contributing dopants. With next generation synchrotron radiation sources and more advanced optics, we foresee that it will be possible to obtain X-ray images of single dopant atoms within resolved radii of 5 nm.

Bayesian Networks (BNs) have become increasingly popular over the last few decades as a tool for reasoning under uncertainty in fields as diverse as medicine, biology, epidemiology, economics and the social sciences. This is especially true in real-world areas where we seek to answer complex questions based on hypothetical evidence to determine actions for intervention. However, determining the graphical structure of a BN remains a major challenge, especially when modelling a problem under causal assumptions. Solutions to this problem include the automated discovery of BN graphs from data, constructing them based on expert knowledge, or a combination of the two. This paper provides a comprehensive review of combinatoric algorithms proposed for learning BN structure from data, describing 74 algorithms including prototypical, well-established and state-of-the-art approaches. The basic approach of each algorithm is described in consistent terms, and the similarities and differences between them highlighted. Methods of evaluating algorithms and their comparative performance are discussed including the consistency of claims made in the literature. Approaches for dealing with data noise in real-world datasets and incorporating expert knowledge into the learning process are also covered.

Studies of ocean surface transport often invoke the "Eulerian-mean hypothesis": that wave-agnostic general circulation models neglecting explicit surface waves effects simulate the Eulerian-mean ocean velocity time-averaged over surface wave oscillations. Acceptance of the Eulerian-mean hypothesis motivates reconstructing the total, Lagrangian-mean surface velocity by adding Stokes drift to model output. Here, we show that the Eulerian-mean hypothesis is inconsistent, because wave-agnostic models cannot accurately simulate the Eulerian-mean velocity if Stokes drift is significant compared to the Eulerian-mean or Lagrangian-mean velocity. We conclude that Stokes drift should not be added to ocean general circulation model velocities. We additionally show the viability of the alternative "Lagrangian-mean hypothesis" using a theoretical argument and by comparing a wave-agnostic global ocean simulation with an explicitly wave-averaged simulation. We find that our wave-agnostic model accurately simulates the Lagrangian-mean velocity even though the Stokes drift is significant., 13 pages, 3 figures

Background In Cyprus, breast cancer (BC) is the first in incidence and second in mortality cancer in women. A national screening programme (NSP), targeting women 50-69 years, was introduced in 2007. The aim of this study is to provide a better understanding of cancer trends. Methods Data from the national population-based Cyprus Cancer Registry on adult women diagnosed with BC between 2004-2017 with follow-up until 2019 were analysed as follows: Joinpoint regression for age-adjusted (overall and by tumor stage at diagnosis - TSD) and age-specific rates ( Results Age-adjusted incidence rate increased from 135.3 (2004) to 153.2 (2017) per 100,000, with an annual percentage change (APC) of 1.1% (95%CI: 0.4-1.9). The greatest increase was in the age groups ≥70 years. A positive time trend was found for localized cancers between 2006-2017, while for all other stages nonsignificant trends were detected. Age-adjusted mortality rate increased from 37.0 (2004) to 50.0 (2019) per 100,000 (APC: 2.7%; 95%CI: 1.9-9.4). Significant increases in mortality rates were detected in the age groups ≥70 years. By TSD, increased rates were found at localised and regional stages, however smaller increases were detected since 2007. NS rates for the most recent period (2014-2017) was 93% for localized, 81% for regional, and 32% for distant and did not significantly improve compared to the previous years. Conclusions Trends in BC incidence continues to increase, especially in the older age groups and for early-stage cancers. As expected, since the introduction of the NSP, the incidence of localised cancers increased whilst the incidence of advanced stage cancer decreased, albeit non-significantly. Survival trends did not change but mortality rates for localised and regional cancers increased at a slower pace. Key messages • The introduction of the national screening programme may have played an important role in the increasing BC incidence trends. • Despite survival rates not improving since the introduction of the national screening programme, mortality rates for early-stage cancers show a less steep increase.

Oceanic eddies play a profound role in mixing tracers such as heat, carbon, and nutrients, thereby regulating regional and global climate. Yet, it remains unclear how global oceanic eddy kinetic energy has evolved over the past few decades. Furthermore, coupled climate model predictions generally fail to resolve oceanic mesoscale dynamics, which could limit their accuracy in simulating future climate change. Here we show a global statistically significant increase of the eddy activity using two independent observational datasets of mesoscale variability, one directly measuring currents and the other from sea surface temperature. Regions characterized by different dynamical processes show distinct evolution in the eddy field. For example, eddy-rich regions such as boundary current extensions and the Antarctic Circumpolar Current show a significant increase of 2% and 5% per decade in eddy activity, respectively. In contrast, most of the regions of observed decrease are found in the tropical oceans. Because eddies play a fundamental role in the ocean transport of heat, momentum, and carbon, our results have far-reaching implications for ocean circulation and climate, and the modelling platforms we use to study future climate change.

Numerous Bayesian Network (BN) structure learning algorithms have been proposed in the literature over the past few decades. Each publication makes an empirical or theoretical case for the algorithm proposed in that publication and results across studies are often inconsistent in their claims about which algorithm is ‘best’. This is partly because there is no agreed evaluation approach to determine their effectiveness. Moreover, each algorithm is based on a set of assumptions, such as complete data and causal sufficiency, and tend to be evaluated with data that conforms to these assumptions, however unrealistic these assumptions may be in the real world. As a result, it is widely accepted that synthetic performance overestimates real performance, although to what degree this may happen remains unknown. This paper investigates the performance of 15 state-of-the-art, well-established, or recent promising structure learning algorithms. We propose a methodology that applies the algorithms to data that incorporates synthetic noise, in an effort to better understand the performance of structure learning algorithms when applied to real data. Each algorithm is tested over multiple case studies, sample sizes, types of noise, and assessed with multiple evaluation criteria. This work involved learning approximately 10,000 graphs with a total structure learning runtime of seven months. In investigating the impact of data noise, we provide the first large scale empirical comparison of BN structure learning algorithms under different assumptions of data noise. The results suggest that traditional synthetic performance may overestimate real-world performance by anywhere between 10% and more than 50%. They also show that while score-based learning is generally superior to constraint-based learning, a higher fitting score does not necessarily imply a more accurate causal graph. The comparisons extend to other outcomes of interest, such as runtime, reliability, and resilience to noise, assessed over both small and large networks, and with both limited and big data. To facilitate comparisons with future studies, we have made all data, raw results, graphs and BN models freely available online.

Some structure learning algorithms have proven to be effective in reconstructing hypothetical Bayesian Network graphs from synthetic data. However, in their mission to maximise a scoring function, many become conservative and minimise edges discovered. While simplicity is desired, the output is often a graph that consists of multiple independent subgraphs that do not enable full propagation of evidence. While this is not a problem in theory, it can be a problem in practice. This article examines a novel unconventional associational heuristic called Saiyan, which returns a directed acyclic graph that enables full propagation of evidence. Associational heuristics are not expected to perform well relative to sophisticated constraint-based and score-based learning approaches. Moreover, forcing the algorithm to connect all data variables implies that the forced edges will not be correct at the rate of those identified unrestrictedly. Still, synthetic and real-world experiments suggest that such a heuristic can be competitive relative to some of the well-established constraint-based, score-based and hybrid learning algorithms.

Circulation in the Southern Ocean is unique. The strong wind stress forcing and buoyancy fluxes, in concert with the lack of continental boundaries, conspire to drive the Antarctic Circumpolar Current replete with an intense eddy field. The effect of Southern Ocean eddies on the ocean circulation is significant – they modulate the momentum balance of the zonal flow, and the meridional transport of tracers and mass. The strength of the eddy field is controlled by a combination of forcing (primarily thought to be wind stress) and intrinsic, chaotic, variability associated with the turbulent flow field itself. Here, we present results from an eddy-permitting ensemble of ocean model simulations to investigate the relative contribution of forced and intrinsic processes in governing the variability of Southern Ocean eddy kinetic energy. We find that variations of the eddy field are mostly random, even on longer (interannual) timescales. Where correlations between the wind stress forcing and the eddy field exist, these interactions are dominated by two distinct timescales – a fast baroclinic instability response; and a multi-year process owing to feedback between bathymetry and the mean flow. These results suggest that understanding Southern Ocean eddy dynamics and its larger-scale impacts requires an ensemble approach to eliminate intrinsic variability, and therefore may not yield robust conclusions from observations alone. Plain Language Summary The Southern Ocean is the most turbulent part of the world’s oceans. This turbulence, often referred to as eddies, is critical to the evolution of the Southern Ocean under climate change. But it’s hard to get information about these eddies, because they occur on small scales in a large ocean basin that is poorly observed. In addition, the observational record is quite short, which makes it more difficult to use these observations to study what controls variations of these eddies. For this reason, we take an eddy-permitting ocean model, and run it 50 times with the same forcing (but a slightly different initial state). The chaotic nature of the turbulent ocean means that these model runs exhibit different evolutions. Then we use these simulations to study which eddy processes occur as a consequence of the chaotic nature of turbulence and which are forced by the external factors that are common to all model runs (such as wind stress). We conclude that monthly-to-interannual fluctuations of the Southern Ocean eddy field are dominated by chaotic processes; but that the forced variability responds to wind on particular timescales that are controlled by the mechanisms that generate ocean turbulence.

Arctic icebergs, unconstrained sea ice floes, oil slicks, mangrove drifters, lost cargo containers, and other flotsam are known to move at 2-4% of the prevailing wind velocity relative to the water, despite vast differences in the material properties, shapes, and sizes of objects. Here, we revisit the roles of density, aspect ratio, and skin and form drag in determining how an object is driven by winds and water currents. Idealized theoretical considerations show that although substantial differences exist for end members of the parameter space (either very thin or thick and very light or dense objects), most realistic cases of floating objects drift at $\approx$3% of the free-stream wind velocity (measured outside an object's surface boundary layer). This relationship, known as a long-standing rule of thumb for the drift of various types of floating objects, arises from the square root of the ratio of the density of air to that of water. We support our theoretical findings with flume experiments using floating objects with a range of densities and shapes., accepted at the Journal of Physical Oceanography

In Bayesian Networks (BNs), the direction of edges is crucial for causal reasoning and inference. However, Markov equivalence class considerations mean it is not always possible to establish edge orientations, which is why many BN structure learning algorithms cannot orientate all edges from purely observational data. Moreover, latent confounders can lead to false positive edges. Relatively few methods have been proposed to address these issues. In this work, we present the hybrid mFGS-BS (majority rule and Fast Greedy equivalence Search with Bayesian Scoring) algorithm for structure learning from discrete data that involves an observational data set and one or more interventional data sets. The algorithm assumes causal insufficiency in the presence of latent variables and produces a Partial Ancestral Graph (PAG). Structure learning relies on a hybrid approach and a novel Bayesian scoring paradigm that calculates the posterior probability of each directed edge being added to the learnt graph. Experimental results based on well-known networks of up to 109 variables and 10 k sample size show that mFGS-BS improves structure learning accuracy relative to the state-of-the-art and it is computationally efficient.

Background COVID-19 pandemic is affecting populations and regions in different ways. In this study, we assess the Potential Years of Life Lost (PYLL) to COVID-19 across different regions. Methods We used age-group and sex-specific weekly COVID-19 deaths (from January to August 2020) from national primary sources of 17 countries from the C-MOR consortium (Australia, Brazil, Cape Verde, Colombia, Cyprus, France, Georgia, Israel, Kazakhstan, Peru, Norway, England & Wales, Scotland, Slovenia, Sweden, Ukraine, and the United States). PYLL were calculated by summing up the numbers of deaths in each age group multiplied by the remaining years of life up to age 80. Age-standardized PYLL rates (per 100,000 population), using the World (WHO 2000-2025) Standard population as the reference population, were estimated to facilitate comparison across countries. Results Countries in South America displayed the highest PYLL rates (567-1,377 PYLL/100,000). Countries in Asia & Middle East (besides Kazakhstan), Australia, and some European countries (Georgia, Norway, and Slovenia) observed Conclusions South America, and males were found to be the most affected by COVID-19. Ongoing monitoring of the COVID-19 mortality impact is essential in order to assess health system performances, control measures, and identify vulnerable populations. Differences in mortality burden among populations will help public health officials in their efforts to minimize the COVID-19 mortality burden on a local, and on a global level. Key messages Up to August 2020, COVID-19 was a cause of premature mortality in all the participating countries, with South America and males to be more affected. The timing of the pandemic, seasonal trends, the control measures enforced, and underlying social conditions are probable explanations for the differences observed among countries.

Background Lung and colorectal cancer are the leading causes of cancer deaths in Cyprus. We describe preliminary analysis of (overall and by stage of diagnosis) age-standardised 5-year net survival (hereafter reported as NS) trends, for these two cancers. Methods We analysed data from the national population-based Cyprus Cancer Registry on adults (age 15+) diagnosed between 2004-17 with lung or colorectal cancer with follow-up until December 2019. We used cohort approach for 2004-08 (time 1-T1), 2009-13 (T2) and period approach for 2014-17 (T3) to estimate NS and 95% confidence interval (CI). Stage at diagnosis was categorised as localised, regional and distant, based on the Surveillance Epidemiology and End Results. The International Cancer Survival Standard age-specific weights were used for standardisation. Results Between 2004-17, all-stages NS increased from 16% (CI:15-17; T1) to 22% (CI:21-23; T3) and from 57% (CI:56-59; T1) to 61% (CI:60-62; T3) for lung and colorectal cancer, respectively. For lung cancer, NS increased from 37% (CI:29-47; T1) to 55% (CI:47-65; T2) and 65% (CI:57-74; T3) for diagnosis at localised stage, from 28% (CI:23-35; T1 and CI:22-35; T2) to 39% (CI:33-46; T3) for regional stage and from 4% (CI:3-6; T1) to 6% (CI:4-8; T2) and 7% (CI:6-10; T3) for distant stage. For colorectal cancer, NS increased from 75% (CI:71-80; T1) to 76% (CI:72-79; T2) and 84% (CI:80-87; T3) for diagnosis at localised stage, from 65% (CI:62-68; T1) to 69% (CI:66-72; T2) and 72% (CI:69-74; T3) for regional stage and from 11% (CI:9-16; T1) to 12% (CI:9-16; T2) and 15% (CI:12-19; T3) for distant stage. Conclusions Over the study period, survival improvements were observed. Late stage at diagnosis seems to be a contributing factor to decreased survival of lung or colorectal cancer patients in Cyprus. Increasing public awareness of cancer risk factors and symptoms and investment in early detection is vital to reduce delay in diagnosis and treatment and to improve survival. Key messages Lung and Colorectal cancer survival continues to improve although at different pace by stage at diagnosis. Any shift to earlier stages at diagnosis brings substantial survival improvements. Survival trends can help both in the formulation of strategies for cancer prevention, screening, and treatment and in the assessment of their effectiveness.

This paper builds on recent developments in Bayesian network (BN) structure learning under the controversial assumption that the input variables are dependent. This assumption can be viewed as a learning constraint geared towards cases where the input variables are known or assumed to be dependent. It addresses the problem of learning multiple disjoint subgraphs that do not enable full propagation of evidence. This problem is highly prevalent in cases where the sample size of the input data is low with respect to the dimensionality of the model, which is often the case when working with real data. The paper presents a novel hybrid structure learning algorithm, called SaiyanH, that addresses this issue. The results show that this constraint helps the algorithm to estimate the number of true edges with higher accuracy compared to the state-of-the-art. Out of the 13 algorithms investigated, the results rank SaiyanH 4th in reconstructing the true DAG, with accuracy scores lower by 8.1% (F1), 10.2% (BSF), and 19.5% (SHD) compared to the top ranked algorithm, and higher by 75.5% (F1), 118% (BSF), and 4.3% (SHD) compared to the bottom ranked algorithm. Overall, the results suggest that the proposed algorithm discovers satisfactorily accurate connected DAGs in cases where other algorithms produce multiple disjoint subgraphs that often underfit the true graph.

This paper presents a parametric study of the collapse performance of two-story residential buildings equipped with a simple spherical concave deformable rolling isolation system cast in high strength concrete. The collapse performance follows the procedures of FEMA P695 with direct evaluation of the spectral shape effects and accounts for uncertainties. The isolation system consists of isolators with fixed dimensional parameters that include a displacement restraint system and have a displacement capacity of 650 mm. The designs considered in the study have properties that are representative of highly ductile reinforced concrete two-story residential buildings consisting of moment frames and/or shear walls designed in Turkey for a lateral force under elastic conditions corresponding to a spectral acceleration of 1 g. This is representative of areas of the highest seismic hazard in Turkey except for areas controlled by near-fault conditions. The simplicity in the design of the structural system and of the isolation system intend to facilitate application of the system without sophisticated analysis. The seismic collapse performance evaluation demonstrates that residential buildings designed by this procedures and using this isolation system have acceptable collapse performance as stipulated in the ASCE/SEI 7-16 standard.

The mesoscale eddy field plays a key role in the mixing and transport of physical and biological properties and redistribute energy budgets in the ocean. Eddy kinetic energy is commonly defined as the kinetic energy of the time-varying component of the velocity field. However, this definition contains all processes that vary in time, including coherent mesoscale eddies, jets, waves, and large-scale motions. The focus of this paper is on the eddy kinetic energy contained in coherent mesoscale eddies. We present a new method to decompose eddy kinetic energy into oceanic processes. The proposed method uses a new eddy-identification algorithm (TrackEddy). This algorithm is based on the premise that the sea level signature of a coherent eddy can be approximated as a Gaussian feature. The eddy Gaussian signature then allows for the calculation of kinetic energy of the eddy field through the geostrophic approximation. TrackEddy has been validated using synthetic sea surface height data, and then used to investigate trends of eddy kinetic energy in the Southern Ocean using Satellite Sea Surface Height anomaly (AVISO+). We detect an increasing trend of eddy kinetic energy associated with mesoscale eddies in the Southern Ocean. This trend is correlated with an increase of the coherent eddy amplitude and the strengthening of wind stress over the last two decades., Comment: 20 pages, 8 figures; TrackEddy available at https://github.com/josuemtzmo/trackeddy

In this work, we study matter in the cores of proto-neutron stars, focusing on the impact of their composition on the stellar structure. We begin by examining the effects of finite temperature (through a fixed entropy per baryon) and lepton fraction on purely nucleonic matter by making use of the DSH (Du, Steiner & Holt) model. We then turn our attention to a relativistic mean-field model containing exotic degrees of freedom, the Chiral Mean Field (CMF) model, again, under the conditions of finite temperature and trapped neutrinos. In the latter, since both hyperons and quarks are found in the cores of large-mass stars, their interplay and the possibility of mixtures of phases is taken into account and analysed. Finally, we discuss how stellar rotation can affect our results.

This paper presents a study for the development and qualification of a practical low-cost seismic isolation system for lightweight residential construction. The study concludes that single concave rolling isolators cast in high strength concrete with a deformable steel-reinforced plastic rolling ball and a displacement restraint system represent a promising isolation system. A full size isolator suitable for application to typical reinforced concrete houses in Turkey was built and tested. The vertical stiffness, creep characteristics under gravity load and the lateral force–displacement characteristics have been studied. Models for behavior have been developed and validated. These models are useful for response history analysis in commonly available commercial and open source analysis software.

Παρουσιάζεται μία θεωρία που αναπτύχθηκε πρόσφατα για το σχηματισμό και την συντήρηση αεροχείμαρρων στις ατμόσφαιρες πλανητών. Η καινοτομία της θεωρίας αυτής είναι ότι αντιμετωπίζει το σχηματισμό και τη συντήρηση των αεροχείμαρρων όχι μέσω της μελέτης της δυναμικής ανεξαρτήτων τροχιών του ρευστού στο θεσεογραφικό χώρο αλλά μέσω της μελέτης της δυναμικής που υπακούει η ίδια η στατιστική των τροχιών αυτών. Αυτό επιτυγχάνεται μέσω της Θεωρίας Στοχαστικής Δομικής Ευστάθειας (Stochastic Structural Stability Theory - S3T) η οποία μελετά το κλειστό δυναμικό σύστημα που περιγράφει τη δυναμική των δύο πρώτων στατιστικών ροπών της πλήρους στατιστικής δυναμικής της ροής, αγνοώντας ή παραμετροποιώντας ροπές τρίτης και ανώτερης τάξεως. Η S3T αποτελεί μια αναλυτική, προγνωστική και ποσοτική θεωρία της τυρβώδους κατάστασης η οποία προκύπτει απευθείας από τις εξισώσεις κινήσεις. Μας παρέχει ένα τρόπο εύρεσης στατιστικών τυρβώδων καταστάσεων ισορροπίας και επίσης τρόπο μελέτης της ευστάθειας αυτών. Αστάθεια της στατιστικής δυναμικής της ροής υπονοεί μετάβαση της τύρβης σε μια κατάσταση. Μέσω αυτού του στατιστικού κλεισίματος (closure) γίνεται μελέτη του σχηματισμού δομών μεγάλης κλίμακας σε βαροτροπική τύρβη σε β-επίπεδο. Μελετώντας τη δυναμικής της στατιστικής κατάστασης μία σειρά από νέα φαινόμενα μπορούν να προβλεφθούν, όπως: η ύπαρξη στατιστικών τυρβώδων καταστάσεων ισορροπίας, η αστάθεια τις ομοιογενούς τυρβώδους κατάστασης ως προς το σχηματισμό ζωνικών ροών, συγχωνεύσεις αεροχείμαρρων ως διακλαδώσεις της στατιστικής δυναμικής και η ύπαρξη λανθανουσών ζωνικών ροών (latent jets). Ενώ όλα αυτά τα φαινόμενα δεν μπορούν να προβλεφθούν με τη μελέτης της δυναμικής ανεξαρτήτων τροχιών του ρευστού στο θεσεογραφικό χώρο, επιδεικνύεται ότι οι προβλέψεις της στατιστικής δυναμικής αντικατοπτρίζονται σε τέτοιες τροχιές της ροής. Ως προς το σχηματισμό των δομών μεγάλης κλίμακας δείχνεται ότι σε κάποιες κρίσιμες τιμές των παραμέτρων, οι οποίες προβλέπονται αναλυτικά, γίνεται μία διακλάδωση και η στατιστικά ομοιογενής τυρβώδης κατάσταση γίνεται ανομοιογενής με την εμφάνιση είτε ζωνικών είτε μη-ζωνικών ροών μεγάλης κλίμακας. Οι μηχανισμοί με τους οποίους οι τάσεις Reynolds από τις τυρβώδεις κινήσεις του υγρού οργανώνονται ούτως ώστε να ενισχύσουν μία απειροστή μέση ροή, οδηγώντας έτσι σε στατιστική αστάθεια, μελετώνται διεξοδικά για διάφορες περιοχές των παραμέτρων (της κλίσης της πλανητικής στροβιλότητας, του ρυθμού ανάλωσης λόγω απόσβεσης και του ρυθμού εισροής ενέργειας στη ροή). Δείχνεται ότι για μικρές και μεσαίες τιμές της κλίσης της πλανητικής στροβιλότητας, β, η ροή ορμής σε κατεύθυνση αντίθετα από αυτήν της βαθμίδας (upgradient momentum flux), η οποία είναι υπεύθυνη για το σχηματισμό των δομών μεγάλης κλίμακας, επάγεται μέσω του μηχανισμού Orr, ενώ για μεγάλες τιμές της β μέσω συντονισμού από τριάδες κυμάτων. Δείχνεται επίσης ότι οι S3T αστάθειες εξισορροπούνται σε πεπερασμένο πλάτος, σε συμφωνία με τις ζωνικές ροές που σχηματίζονται σε αριθμητικές προσομοιώσεις των εξισώσεων κίνησης. Διερευνάται επίσης η σχέση μεταξύ του σχηματισμού δομών μεγάλης κλίμακας μέσω της διαμορφωτικής αστάθειας (modulational instability) και της S3T αστάθειας της ομοιογενούς τυρβώδους κατάστασης και δείχνεται ότι η διαμορφωτική αστάθεια δεν είναι παρά μία ειδική περίπτωση της S3T αστάθειας. Τέλος, γίνεται μελέτη της S3T ευστάθειας ανομοιογενών στατιστικών τυρβώδων καταστάσεων στις οποίες υπάρχουν ζωνικές ροές μεγάλης κλίμακας και διερευνάται η σχέση των S3T ασταθειών που προκύπτουν με το φαινόμενο συγχώνευσης αεροχείμαρρων. Αναπτύσσονται μέθοδοι εύρεσης ανομοιογενών στατιστικών τυρβώδων καταστάσεων καθώς επίσης και μελέτης της ευστάθειας αυτών.

Seismic isolation is generally considered an effective earthquake protection strategy. As application of seismic isolation increases, decisions on the use of one particular isolator versus another isolator increasingly depend on computed responses with complex analytical models. Accordingly, validation of analytical models to predict primary (structural) and secondary system (non-structural component) response in seismically isolated buildings becomes very important. This paper presents comparisons of experimental and analytical results on the primary and secondary system response of a building model in order to provide information on the accuracy of the predicted response. The tested model was configured as a 6-story building at quarter length scale in a moment-frame configuration, and with the following seismic isolation systems: a) Low damping elastomeric bearings with and without linear or nonlinear viscous dampers, b) Single Friction Pendulum (FP) bearings with and without linear or nonlinear viscous dampers, and c) Lead-rubber bearings. Response quantities compared include story drifts and isolator shear forces and displacements for the primary system, and peak floor total velocities and floor response spectra that relate to secondary system response. This paper presents samples results out of a total of 288 comparisons of experimental and analytical results presented in an MCEER report. It is shown that the primary and secondary system response is computed with sufficient accuracy by the analytical models but some response quantities may be underestimated or overestimated by significant amounts.

This study aimed to evaluate the use of a by-product, olive cake silage (OCS), as a forage replacement in sheep diets for the improvement of fatty acid (FA) content of milk and thus, the lipids of the ovine halloumi cheese produced. Sixty second-parity purebred Chios ewes in mid-lactation were assigned to three diet treatments (2 lots of 10 animals per treatment) receiving 0%, 10%, and 20% of OCS on dry matter basis for 3 weeks (treatments S0, S10, and S20, respectively). Halloumi cheese was manufactured from fresh raw milk of ewes fed the three different diets. Inclusion of OCS in the diets increased linearly the concentration in milk of unsaturated FA up to 20%, monounsaturated FA up to 23%, polyunsaturated FA up to 11%, rumenic acid (CLA cis-9, trans-11) up to 61%, and consequently reduced the atherogenicity and thrombogenicity milk indices by 31% and 27%, for the S10 and S20 treatments, respectively, compared with the control treatment. Moreover, these differences were carried over to the lipid profile of ovine halloumi cheese showing, on average, more than 25% increase of unsaturated, polyunsaturated, and monounsaturated FA, with particularly enhanced oleic and rumenic acid content. These changes resulted in reduced atherogenicity by 29% and 45% and thrombogenicity by 23% and 24% of ovine halloumi cheese made from milk of S10 and S20 diets, respectively. Milk yield, milk fat, or protein content was not affected by S10 or S20 feeding treatments compared to control. Overall, the applied ensiling method of olive cake produces a by-product that can be included as a forage replacement up to 20% of DM intake in Chios sheep without adversely affecting the lactating performance. Furthermore, the present study showed that such substitution improves the lipid quality of milk and related halloumi cheese enriching these ovine dairy products with beneficial to human health fatty acids.

Despite the nonlinear nature of turbulence, there is evidence that part of the energy-transfer mechanisms sustaining wall turbulence can be ascribed to linear processes. The different scenarios stem from linear stability theory and comprise exponential instabilities, neutral modes, transient growth from non-normal operators, and parametric instabilities from temporal mean-flow variations, among others. These mechanisms, each potentially capable of leading to the observed turbulence structure, are rooted in theoretical and conceptual arguments. Whether the flow follows any or a combination of them remains elusive. Here, we evaluate the linear mechanisms responsible for the energy transfer from the streamwise-averaged mean-flow ($\bf U$) to the fluctuating velocities ($\bf u'$). We use cause-and-effect analysis based on interventions. This is achieved by direct numerical simulation of turbulent channel flows at low Reynolds number, in which the energy transfer from $\bf U$ to $\bf u'$ is constrained to preclude a targeted linear mechanism. We show that transient growth is sufficient for sustaining realistic wall turbulence. Self-sustaining turbulence persists when exponential instabilities, neutral modes, and parametric instabilities of the mean flow are suppressed. We further show that a key component of transient growth is the Orr/push-over mechanism induced by spanwise variations of the base flow. Finally, we demonstrate that an ensemble of simulations with various frozen-in-time $\bf U$ arranged so that only transient growth is active, can faithfully represent the energy transfer from $\bf U$ to $\bf u'$ as in realistic turbulence. Our approach provides direct cause-and-effect evaluation of the linear energy-injection mechanisms from $\bf U$ to $\bf u'$ in the fully nonlinear system and simplifies the conceptual model of self-sustaining wall turbulence.

Causal Bayesian networks have become a powerful technology for reasoning under uncertainty in areas that require transparency and explainability, by relying on causal assumptions that enable us to simulate hypothetical interventions. The graphical structure of such models can be estimated by structure learning algorithms, domain knowledge, or a combination of both. Various knowledge approaches have been proposed in the literature that enables us to specify prior knowledge that constrains or guides these algorithms. This paper introduces some novel, and also describes some existing, knowledge-based approaches that enable us to combine structure learning with knowledge obtained from heterogeneous sources. We investigate the impact of these approaches on structure learning across different algorithms, case studies and settings that we might encounter in practice. Each approach is assessed in terms of effectiveness and efficiency, including graphical accuracy, model fitting, complexity, and runtime; making this the first paper that provides a comparative evaluation of a wide range of knowledge approaches for structure learning. Because the value of knowledge depends on what data are available, we illustrate the results both with limited and big data. While the overall results show that knowledge becomes less important with big data due to higher learning accuracy rendering knowledge less important, some of the knowledge approaches are found to be more important with big data. Amongst the main conclusions is the observation that reduced search space obtained from knowledge does not always imply reduced computational complexity, perhaps because the relationships implied by the data and knowledge are in tension.

In remote sensing applications, constraints of power, processing, and communications limit information acquisition. Pre-training the IoT improves the performance in information acquisition tasks such as detection, classification, and estimation. However, light and inexpensive IoT hardware still need to operate with strict resource constraints. In this paper, we provide a method for modeling the IoT operating characteristics that link information acquisition performance to resource use. The goal of modeling is to improve understanding of how to optimally utilize constrained resources to improve information acquisition performance. The proposed method is demonstrated using field, simulation, and lab-based experiments with real data and practical hardware for an underwater sound classification application utilizing deep learning.

In this work, we investigate how the assumption of chemical equilibrium with leptons affects the deconfinement phase transition to quark matter. This is done within the framework of the Chiral Mean Field model (CMF) allowing for non-zero net strangeness, corresponding to the conditions found in astrophysical scenarios. We build 3-dimensional QCD phase diagrams with temperature, baryon chemical potential, and either charge or isospin fraction or chemical potential to show how the deconfinement region collapses to a line in the special case of chemical equilibrium, such as the one established the interior of cold catalyzed neutron stars.

In this work, we examine in detail the difference between constraining the electric charge fraction and isospin fraction when calculating the deconfinement phase transition in the presence of net strangeness. We present relations among charge and isospin fractions and the corresponding chemical potentials and draw 3-dimensional QCD phase diagrams for matter out of weak equilibrium. Finally, we briefly discuss how our results can be applied to comparisons of matter created in heavy ion collisions and binary neutron star mergers., arXiv admin note: text overlap with arXiv:2004.03039

The influence of thermal fluctuations on fermion pairing is investigated using a semiclassical treatment of fluctuations. When the average pairing gaps along with those differing by one standard deviation are used, the characteristic discontinuity of the specific heat at the critical temperature Tc in the BCS formalism with its most probable gap becomes smooth. This indicates the suppression of a second order phase transition as experimentally observed in nanoparticles and several nuclei. Illustrative calculations using the constant spacing model and the recently introduced random spacing model are presented.

Score-based algorithms that learn Bayesian Network (BN) structures provide solutions ranging from different levels of approximate learning to exact learning. Approximate solutions exist because exact learning is generally not applicable to networks of moderate or higher complexity. In general, approximate solutions tend to sacrifice accuracy for speed, where the aim is to minimise the loss in accuracy and maximise the gain in speed. While some approximate algorithms are optimised to handle thousands of variables, these algorithms may still be unable to learn such high dimensional structures. Some of the most efficient score-based algorithms cast the structure learning problem as a combinatorial optimisation of candidate parent sets. This paper explores a strategy towards pruning the size of candidate parent sets, and which could form part of existing score-based algorithms as an additional pruning phase aimed at high dimensionality problems. The results illustrate how different levels of pruning affect the learning speed relative to the loss in accuracy in terms of model fitting, and show that aggressive pruning may be required to produce approximate solutions for high complexity problems.

This study investigates the effect of displacement restraint on the collapse performance of seismically isolated buildings. The displacement restraints considered are stiffening triple Friction Pendulum (FP) bearings and moat walls. The study is based on 6-story perimeter frame seismically isolated buildings designed with special concentrically braced frames (SCBF) and special moment resisting frames (SMF) for a location in California using the minimum criteria of ASCE/SEI 7-10 and ASCE/SEI 7-16 and also using enhanced designs. Results from pushover analysis and nonlinear response history analysis demonstrate that proper designs require a balance between the value of RI for the design of the superstructure and the displacement capacity of the isolators. The paper shows that isolation systems with sufficient displacement capacity before engaging the displacement restraint and a RI factor consistent to that displacement capacity may have an acceptable collapse risk. Values of the design parameters for the superstructure and the displacement capacity and behavior of the isolation system for achieving acceptable collapse risk are presented.

This article presents an analytical study of the seismic collapse performance of seismically isolated buildings and comparable non-isolated buildings. The study is based on archetypical 6-story perimeter frame seismically isolated buildings designed with special concentrically braced frames (SCBF), ordinary concentrically braced frames (OCBF) and special moment resisting frames (SMF) for a location in California using the minimum criteria of ASCE/SEI 7-10 and ASCE/SEI 7-16 and also using a number of enhanced designs. The isolation system consists of triple Friction Pendulum (FP) isolators with stiffening behavior at large displacement. Additionally, double concave sliding isolators are considered and designed per minimum criteria of ASCE/SEI 7 and without a displacement restrainer, a practice permitted by the standards. Non-isolated structures, also with braced and moment frame configurations, are designed using the minimum criteria of ASCE/SEI 7 and studied. The study concludes that seismically isolated buildings designed by the minimum criteria of ASCE/SEI 7 of either 2010 or 2016 may have unacceptable probability of collapse in the Risk-Targeted Maximum Considered Earthquake (MCER). The probability of collapse in the MCER becomes acceptable when they are designed with enhanced criteria of RI=1.0 and with isolators having a displacement capacity at initiation of stiffening equal to 1.5 times the demand in the MCER. It is also observed that designs that meet the minimum criteria of ASCE/SEI 7 of either 2010 or 2016 and without any displacement restrainer have unacceptably high probabilities of collapse.

While decision trees are a popular formal and quantitative method for determining an optimal decision from a finite set of choices, for all but very simple problems they are computationally intractable. For this reason, Influence Diagrams (IDs) have been used as a more compact and efficient alternative. However, most algorithmic solutions assume that all chance variables are discrete, whereas in practice many are continuous. For such ‘Hybrid’ IDs (HIDs) the current-state-of-the-art algorithms suffer from various limitations on the kinds of inference that can be performed. This paper presents a novel method that overcomes a number of these limitations. The method solves a HID by transforming it to a Hybrid Bayesian Network (HBN) and carrying out inference on this HBN using Dynamic Discretization (DD). It generates a simplified decision tree from the propagated HBN to compute and present the optimal decisions under different decision scenarios. To provide satisfactory performance the method uses ‘inconsistent evidence’ to model functional and structural asymmetry. By using the entire marginal probability distribution of the continuous utility and chance nodes, rather than expected values alone, our method also enhances decision analysis by offering the possibility to consider additional statistics other than expected utility, such as measures of risk. We illustrate our method by using the oil wildcatter example and its variations with continuous nodes. We also use a financial score to combine risk and return measures, for illustration.

The electric power network is considered to be critical infrastructure. Electric utility companies and electric power administrators are increasingly interested in strengthening the security and resilience of their networks against earthquakes. Important components of these networks are high-voltage electrical transformers. Earlier studies in the use of seismic isolation systems for electrical transformers and other equipment produced results that clearly show the benefits of seismic isolation in the horizontal direction. However, vertical ground motions were transmitted through the isolation system unchanged or even magnified. In response to this problem, this paper presents the development and validation of a compact, effective, reliable and cost-effective combined horizontal–vertical seismic isolation system for use in electrical equipment.

In decision theory models, expected value of partial perfect information (EVPPI) is an important analysis technique that is used to identify the value of acquiring further information on individual variables. EVPPI can be used to prioritize the parts of a model that should be improved or identify the parts where acquiring additional data or expert knowledge is most beneficial. Calculating EVPPI of continuous variables is challenging, and several sampling and approximation techniques have been proposed. This paper proposes a novel approach for calculating EVPPI in hybrid influence diagram (HID) models (these are influence diagrams (IDs) containing both discrete and continuous nodes). The proposed approach transforms the HID into a hybrid Bayesian network and makes use of the dynamic discretization and the junction tree algorithms to calculate the EVPPI. This is an approximate solution (no feasible exact solution is possible generally for HIDs) but we demonstrate it accurately calculates the EVPPI values. Moreover, unlike the previously proposed simulation-based EVPPI methods, our approach eliminates the requirement of manually determining the sample size and assessing convergence. Hence, it can be used by decision-makers who do not have deep understanding of programming languages and sampling techniques. We compare our approach to the previously proposed techniques based on two case studies.

Seismic isolation is generally considered an effective earthquake protection strategy. As application of seismic isolation increases, decisions on the use of one particular isolator versus another isolator increasingly depend on computed responses with complex analytical models. Accordingly, validation of analytical models to predict primary (structural) and secondary system (non-structural component) response in seismically isolated buildings becomes very important. This paper presents comparisons of experimental and analytical results on the primary and secondary system response of a building model in order to provide information on the accuracy of the predicted response. The tested model was configured as a 6-story building at quarter length scale in a moment-frame configuration, and with the following seismic isolation systems: a) Low damping elastomeric bearings with and without linear or nonlinear viscous dampers, b) Single Friction Pendulum (FP) bearings with and without linear or nonlinear viscous dampers, and c) Lead-rubber bearings. Response quantities compared include story drifts and isolator shear forces and displacements for the primary system, and peak floor total velocities and floor response spectra that relate to secondary system response. This paper presents samples results out of a total of 288 comparisons of experimental and analytical results presented in an MCEER report. It is shown that the primary and secondary system response is computed with sufficient accuracy by the analytical models but some response quantities may be underestimated or overestimated by significant amounts.

This study presents a probabilistic response analysis of seismically isolated electrical transformers with emphasis on comparing the performance of equipment that are non-isolated to equipment that are isolated only in the horizontal direction or are isolated by a three-dimensional isolation system. The performance is assessed by calculating the probability of failure as function of the seismic intensity with due consideration of: (a) horizontal and vertical ground seismic motions, (b) displacement capacity of the seismic isolation system, (c) limit states of electrical bushings, (d) details of construction of the isolation system, (e) weight of the isolated transformer, and (f) bushing geometry and configuration. Calculations of the probability of failure within the lifetime of isolated and non-isolated transformers at selected locations are also performed. The results of this study demonstrate that seismic isolation systems can improve the seismic performance for a wide range of parameters and that systems which isolate in both the horizontal and vertical directions can be further effective. The seismic assessment methodology presented can be used for: (a) deciding on the need to use seismic isolation and selecting the properties of the isolation system for transformers depending on the design limits, location, and configuration of transformer and (b) calculating the mean annual frequency of functional failure and the corresponding probability of failure over the lifetime of the equipment. The results may also be used to assess the seismic performance of electric transmission networks under scenarios of component failures.