Discrete Bins Research Articles

Back to the future: the persistent problems of hybrid war

Abstract What has happened to the spectre of Russian hybrid warfare that once haunted strategists, policy-makers and scholars? Many of the characteristics of Russia's 2014 invasion of Crimea have been absent in Moscow's current war, prompting some to wonder whether Russia has moved beyond ‘hybrid warfare’ as it wages a brutal conventional war in Ukraine. Scholars and practitioners should not pack up their hybrid war handbooks just yet, however. The need to limit conflict escalation between nuclear-armed adversaries suggests that non-military asymmetric tools of will increasingly define great power competition in the twenty-first century. Most urgently, scholars must bring some order to the conceptual debate over what is and is not hybrid war. In this policy paper, we propose a classification scheme that can be applied irrespective of the type of conflict, one that illuminates the wide range of political, social, economic and military weapons states may use in future conflicts. This approach not only allows us to move beyond endless debates over which tactics belong in the definition of hybrid war, but it also allows us to describe changes over time in how countries wage multi-domain wars. By describing the ‘hybridity’ of all conflict along three continuous axes of scale, militarization and centralization, scholars and policy-makers alike can avoid dangerous blind spots that are likely to arise when we try to categorize conflicts into discrete typological bins.

Learning Offset Probability Distribution for Accurate Object Detection

Object detection combines object classification and object localization problems. Current object detection methods heavily depend on regression networks to locate objects, which are optimized with various regression loss functions to predict offsets between candidate boxes and objects. However, these regression losses are difficult to assign the appropriate penalties for samples with large offset errors, resulting in suboptimal regression networks and inaccurate object offsets. In this article, we consider object location as offset bin classification problem, and propose a distance-aware offset bin classification network optimized with multiple binary cross entropy losses to learn various offset probability distribution, including single label distribution and distance-aware label distribution. On one hand, it provides gradient contributions for different samples based on the bounded probability instead of previous incalculable offset error. On the other hand, it explores the distance correlations between discrete offset bins to facilitate network learning. Specifically, we discretize the continuous offset into a number of bins, and predict the probability of each offset bin, in which the probability should be higher for the offset bin closer to the target offsets, and vice versa. Furthermore, we propose an expectation-based offset prediction and a hierarchical focusing method to improve the precision of prediction. We conduct extensive experiments to evaluate the effectiveness of our method. In addition, our method can be conveniently and flexibly inserted into existing object detection methods, which consistently achieves a large gain based on popular anchor-based and anchor-free methods on the PASCAL VOC, MS-COCO, KITTI, and CrowdHuman datasets. Code will be released at: https://github.com/QiuHeqian/DBC .

BinsFormer: Revisiting Adaptive Bins for Monocular Depth Estimation.

Monocular depth estimation (MDE) is a fundamental task in computer vision and has drawn increasing attention. Recently, some methods reformulate it as a classification-regression task to boost the model performance, where continuous depth is estimated via a linear combination of predicted probability distributions and discrete bins. In this paper, we present a novel framework called BinsFormer, tailored for the classification-regression-based depth estimation. It mainly focuses on two crucial components in the specific task: 1) proper generation of adaptive bins and 2) sufficient interaction between probability distribution and bins predictions. To specify, we employ a Transformer decoder to generate bins, novelly viewing it as a direct set-to-set prediction problem. We further integrate a multi-scale decoder structure to achieve a comprehensive understanding of spatial geometry information and estimate depth maps in a coarse-to-fine manner. Moreover, an extra scene understanding query is proposed to improve the estimation accuracy, which turns out that models can implicitly learn useful information from the auxiliary environment classification task. Extensive experiments on the KITTI, NYU, and SUN RGB-D datasets demonstrate that BinsFormer surpasses state-of-the-art MDE methods with prominent margins. Code and pretrained models are made publicly available at https://github.com/zhyever/Monocular-Depth-Estimation-Toolbox/tree/ main/configs/binsformer.

A continuous form of the Perez diffuse sky model for forward and reverse transposition

The Perez 1990 transposition model has emerged as the preferred choice for estimating global tilted irradiance, also known as plane-of-array irradiance. One notable drawback is the Perez model’s reliance on empirical coefficients assigned to discrete bins of the sky clearness parameter, resulting in discontinuities in the calculated tilted irradiance. In this study, we present a novel method to eliminate the discontinuities of the Perez model by replacing the empirical look-up table with a set of six quadratic splines. This is facilitated by transforming the unbounded sky clearness parameter (epsilon) to an equivalent bounded parameter (zeta).Transposition using the original Perez model and continuous Perez–Driesse model are compared for multiple orientations at two locations. The two models produce very similar deviation statistics, meaning the continuous version can be used as a plug-in replacement for the original. Reverse transposition is demonstrated using the Perez–Driesse model together with a new continuous version of the Erbs diffuse fraction model and a simple bisection solution search. This combination achieves a substantially higher success rate than the GTI-DIRINT algorithm in our tests.

Cardiovascular variability, sociodemographics, and biomarkers of disease: the MIDUS study.

Introduction: Like heart rate, blood pressure (BP) is not steady but varies over intervals as long as months to as short as consecutive cardiac cycles. This blood pressure variability (BPV) consists of regularly occurring oscillations as well as less well-organized changes and typically is computed as the standard deviation of multiple clinic visit-to-visit (VVV-BP) measures or from 24-h ambulatory BP recordings (ABPV). BP also varies on a beat-to-beat basis, quantified by methods that parse variation into discrete bins, e.g., low frequency (0.04-0.15Hz, LF). However, beat-to-beat BPV requires continuous recordings that are not easily acquired. As a result, we know little about the relationship between LF-BPV and basic sociodemographic characteristics such as age, sex, and race and clinical conditions. Methods: We computed LF-BPV during an 11-min resting period in 2,118 participants in the Midlife in the US (MIDUS) study. Results: LF-BPV was negatively associated with age, greater in men than women, and unrelated to race or socioeconomic status. It was greater in participants with hypertension but unrelated to hyperlipidemia, hypertriglyceridemia, diabetes, elevated CRP, or obesity. LF-diastolic BPV (DBPV), but not-systolic BPV (SBPV), was negatively correlated with IL-6 and s-ICAM and positively correlated with urinary epinephrine and cortisol. Finally, LF-DBPV was negatively associated with mortality, an effect was rendered nonsignificant by adjustment by age but not other sociodemographic characteristics. Discussion: These findings, the first from a large, national sample, suggest that LF-BPV differs significantly from VVV-BP and ABPV. Confirming its relationship to sociodemographic risk factors and clinical outcomes requires further study with large and representative samples.

Photon-counting computed tomography thermometry via material decomposition and machine learning

Thermal ablation procedures, such as high intensity focused ultrasound and radiofrequency ablation, are often used to eliminate tumors by minimally invasively heating a focal region. For this task, real-time 3D temperature visualization is key to target the diseased tissues while minimizing damage to the surroundings. Current computed tomography (CT) thermometry is based on energy-integrated CT, tissue-specific experimental data, and linear relationships between attenuation and temperature. In this paper, we develop a novel approach using photon-counting CT for material decomposition and a neural network to predict temperature based on thermal characteristics of base materials and spectral tomographic measurements of a volume of interest. In our feasibility study, distilled water, 50 mmol/L CaCl2, and 600 mmol/L CaCl2 are chosen as the base materials. Their attenuations are measured in four discrete energy bins at various temperatures. The neural network trained on the experimental data achieves a mean absolute error of 3.97 °C and 1.80 °C on 300 mmol/L CaCl2 and a milk-based protein shake respectively. These experimental results indicate that our approach is promising for handling non-linear thermal properties for materials that are similar or dissimilar to our base materials.

Modelling grain-size distributions in C-type shocks using a discrete power-law model

In this paper we discuss the implementation of a discrete, piecewise power-law grain-size distribution method into a numerical multifluid MHD code as described in Sumpter and Van Loo (2020). Such a description allows to capture the full size range of dust grains and their dynamical effects. The only assumptions are that grains within a single discrete bin have the same velocity and charge. We test the implementation by modelling plane-parallel C-type shocks and compare the results with shock models of multispecies grain models. We find that both the discrete and multispecies grain models converge to the same shock profile. However, the convergence for the discrete models is faster than for the multispecies grain models. For the pure advection models a single discrete bin is sufficient, while the multispecies grain models need a minimum of 8 grain species. When including grain sputtering the necessary number of discrete bins increases to 4, as the grain distribution cannot be described by a single power-law as in the advection models. The multispecies grain models still need more grain species to model the distribution, but the number does not increase compared to the pure advection models. Our results show that modelling the grain distribution function using a discrete distribution reduces the computational cost needed to capture the grain physics significantly.

A parametric and feasibility study for data sampling of the dynamic mode decomposition: Spectral insights and further explorations

The present work extends the parametric investigation on the sampling nuances of dynamic mode decomposition (DMD) under Koopman analysis. Through turbulent wakes, the study corroborated the generality of universal convergence states for all DMD implementations. It discovered implications of sampling range and resolution—determinants of spectral discretization by discrete bins and the highest resolved frequency range, respectively. The work reaffirmed the necessity of the convergence state for sampling independence, too. Results also suggested that the observables derived from the same flow may contain dynamically distinct information, thus altering the DMD output. Surface pressure and vortex fields are optimal for characterizing the structure and the flow field, respectively. Pressure, velocity magnitude, and turbulence kinetic energy also suffice for general applications, but Reynolds stresses and velocity components shall be avoided. Mean-subtraction is recommended for the best approximations of Koopman eigen tuples. Furthermore, the parametric investigation on truncation discovered some low-energy states that dictate a system's temporal integrity. The best practice for order reduction is to avoid truncation and employ dominant mode selection on a full-state subspace, though large-degree truncation supports fair data reconstruction with low computational cost. Finally, this work demonstrated synthetic noise resulting from pre-decomposition interpolation. In unavoidable interpolations to increase the spatial dimension n, high-order schemes are recommended for better retention of original dynamics. Finally, the observations herein, derived from inhomogeneous anisotropic turbulence, offer constructive references for DMD on fluid systems, if not also for others beyond fluid mechanics.

Non-Gaussian photonic state engineering with the quantum frequency processor

Non-Gaussian quantum states of light are critical resources for optical quantum information processing, but methods to generate them efficiently remain challenging to implement. Here we introduce a generic approach for non-Gaussian state production from input states populating discrete frequency bins. Based on controllable unitary operations with a quantum frequency processor, followed by photon-number-resolved detection of ancilla modes, our method combines recent developments in both frequency-based quantum information and non-Gaussian state preparation. Leveraging and refining the K-function representation of quantum states in the coherent basis, we develop a theoretical model amenable to numerical optimization and, as specific examples, design quantum frequency processor circuits for the production of Schr\"{o}dinger cat states, exploring the performance tradeoffs for several combinations of ancilla modes and circuit depth. Our scheme provides a valuable general framework for producing complex quantum states in frequency bins, paving the way for single-spatial-mode, fiber-optic-compatible non-Gaussian resource states.

Dosimetric Impact of Deformable Image Registration-Based Target Volume Propagation in Stereotactic Body Radiotherapy of Ultracentral Lung Tumors

In stereotactic body radiotherapy (SBRT) of non-small cell lung cancer (NSCLC), the use of steep dose gradients and reduced margins has improved local control while limiting toxicity. Clinical benefit of these techniques is dependent on accurate target definition. NSCLC patients are vulnerable to errors in target definition because lung tumors deform with respiratory motion. For ultracentral lung tumors, phantom studies suggest that auto-propagation of internal gross tumor volumes (IGTVs) can be inaccurate. In SBRT of ultracentral tumors, deformable image registration (DIR) is widely used for target volume propagation. We hypothesized that errors in DIR-based IGTV propagation could jeopardize SBRT target coverage for ultracentral tumors despite use of population-based safety margins.Nine patients with ultracentral NSCLC tumors appropriate for SBRT were studied. For each patient, (1) 4D CT was acquired into 10 discrete phase-based bins; (2) the GTV was manually segmented on all phases; and (3) DIR-based IGTV propagation was carried out for each manual GTV on every phase, resulting in 10 propagated IGTVs (P-IGTVs) per patient and 90 for the entire cohort. Ground truth IGTV (GT-IGTV) was defined as the union of all manual GTVs from all 10 phases. Dice coefficient was used to measure the similarity of P-IGTV relative to GT-IGTV. For each patient, SBRT plans were generated for P-IGTVs with greatest (Plan A) and least (Plan B) Dice coefficients using a PTV margin of 5 mm. Prescriptions were 60 Gy in 8 fractions with V60 Gy > 95% PTV. Dose constraints for organs-at-risk (OARs) were defined according to published prospective trials. For each patient, dosimetric differences between Plans A and B for the ground truth PTV (GT-IGTV + 5 mm margin) and OARs were examined.All tumors (6 in the right lung, 3 in the left) were within 2 cm of the proximal bronchial tree and at hilar level. Table 1 compares IGTVs and target coverage with median and range values. Plans A and B differed significantly in ground truth PTV (GT-PTV) coverage at both 99% and 95% of Rx dose (Kruskal-Wallis tests; P = 0.003). There were no significant differences between plans for dose to OARs. For Plan A, 7 out 9 plans were generated from GTVs at near-end of inhalation (respiratory phases 8 or 9); for Plan B, 7/9 plans were generated from GTVs at end of inhalation through mid-exhalation (phases 0 through 3).In SBRT of ultracentral tumors, patterns of error exist in DIR-based IGTV propagation that can compromise target coverage despite the use of population-based safety margins. These findings are important to enhance efforts for more robust and accurate targeting in lung SBRT.

The hidden WASH needs of perimenopausal women

Abstract Meeting the needs of an ageing population, particularly in the global South, is increasingly an issue for WASH service provision. An ageing-related issue for women, warranting specific attention from the WASH sector, is the perimenopause, but this is often not discussed publicly. Drawing on data from the UK and Ghana, this paper provides recommendations for meeting the additional WASH needs of women as they experience the perimenopause (the transition period to menopause). Finding these ‘hidden’ needs involved a UK-based phenomenological review and mixed feminist methods in two low-income urban communities in Accra and Kumasi, Ghana. The hidden WASH needs of perimenopausal women included understanding the perimenopause, menstrual hygiene management, urine incontinence management, bathing, and laundry. Community awareness, engagement and empowerment, and ensuring good perimenopausal health can begin to address these needs. Infrastructural measures for perimenopausal women include a continuous water supply, user-friendly bathing and laundry infrastructure, and gender-sensitive, accessible toilets with discrete sanitary disposal bins. High-absorbency sanitary products are important for managing heavy menstruation. A better understanding of the needs of perimenopausal women and training on how to meet these would benefit the WASH sector in ensuring that perimenopausal women are not left behind in efforts to meet the Sustainable Development Goals.

Does pericentral mu-rhythm “power” corticomotor excitability? – A matter of EEG perspective

BackgroundElectroencephalography (EEG) and single-pulse transcranial magnetic stimulation (spTMS) of the primary motor hand area (M1-HAND) have been combined to explore whether the instantaneous expression of pericentral mu-rhythm drives fluctuations in corticomotor excitability, but this line of research has yielded diverging results. ObjectivesTo re-assess the relationship between the mu-rhythm power expressed in left pericentral cortex and the amplitude of motor potentials (MEP) evoked with spTMS in left M1-HAND. Methods15 non-preselected healthy young participants received spTMS to the motor hot spot of left M1-HAND. Regional expression of mu-rhythm was estimated online based on a radial source at motor hotspot and informed the timing of spTMS which was applied either during epochs belonging to the highest or lowest quartile of regionally expressed mu-power. Using MEP amplitude as dependent variable, we computed a linear mixed-effects model, which included mu-power and mu-phase at the time of stimulation and the inter-stimulus interval (ISI) as fixed effects and subject as a random effect. Mu-phase was estimated by post-hoc sorting of trials into four discrete phase bins. We performed a follow-up analysis on the same EEG-triggered MEP data set in which we isolated mu-power at the sensor level using a Laplacian montage centered on the electrode above the M1-HAND. ResultsPericentral mu-power traced as radial source at motor hot spot did not significantly modulate the MEP, but mu-power determined by the surface Laplacian did, showing a positive relation between mu-power and MEP amplitude. In neither case, there was an effect of mu-phase on MEP amplitude. ConclusionThe relationship between cortical oscillatory activity and cortical excitability is complex and minor differences in the methodological choices may critically affect sensitivity.

Multilayer Connector Hub Mapping Reveals Key Brain Regions Supporting Expressive Language.

Introduction: How components of the distributed brain networks that support cognition participate in typical functioning remains a largely unanswered question. An important subgroup of regions in the larger network are connector hubs, which are areas that are highly connected to several other functionally specialized sets of regions, and are likely important for sensorimotor integration. The present study attempts to characterize connector hubs involved in typical expressive language functioning using a data-driven, multimodal, full multilayer magnetoencephalography (MEG) connectivity-based pipeline.Methods: Twelve adolescents, 16–18 years of age (five males), participated in this study. Participants underwent MEG scanning during a verb generation task. MEG and structural connectivity were calculated at the whole-brain level. Amplitude/amplitude coupling (AAC) was used to compute functional connections both within and between discrete frequency bins. AAC values were then multiplied by a binary structural connectivity matrix, and then entered into full multilayer network analysis. Initially, hubs were defined based on multilayer versatility and subsequently reranked by a novel measure called delta centrality on interconnectedness (DCI). DCI is defined as the percent change in interfrequency interconnectedness after removal of a hub.Results: We resolved regions that are important for between-frequency communication among other areas during expressive language, with several potential theoretical and clinical applications that can be generalized to other cognitive domains.Conclusion: Our multilayer, data-driven framework captures nonlinear connections that span across scales that are often missed in conventional analyses. The present study suggests that crucial hubs may be conduits for interfrequency communication between action and perception systems that are crucial for typical functioning.Impact statementWe present methodology to characterize regions supporting cross-frequency communication in the distributed language network. There are 3 key innovations: (1) incorporation of a structural connectivity constraint based on diffusion magnetic resonance imaging (MRI), (2) use of a full multilayer framework that captures both within- and between-frequency connections, and (3) introduction of a new metric, delta centrality on interconnectedness (DCI), that quantifies the importance of a region for cross-frequency coupling.

Using hidden Markov models to find discrete targets in continuous sociophonetic data

AbstractAdvances in sociophonetic research resulted in features once sorted into discrete bins now being measured continuously. This has implied a shift in what sociolinguists view as the abstract representation of the sociolinguistic variable. When measured discretely, variation is variation in selection: one variant is selected for production, and factors influencing language variation and change are influencing the frequency at which variants are selected. Measured continuously, variation is variation in execution: speakers have a single target for production, which they approximate with varying success. This paper suggests that both approaches can and should be considered in sociophonetic analysis. To that end, I offer the use of hidden Markov models (HMMs) as a novel approach to find speakers’ multiple targets within continuous data. Using thelotvowel among whites in Greater St. Louis as a case study, I compare 2-state and 1-state HMMs constructed at the individual speaker level. Ten of fifty-two speakers’ production is shown to involve the regular use of distinct fronted and backed variants of the vowel. This finding illustrates HMMs’ capacity to allow us to consider variation as both variant selection and execution, making them a useful tool in the analysis of sociophonetic data.

Low-complexity GLRT for FDA radar without training data

Conventionally adaptive detection for frequency diverse array (FDA) radar requires extra training data to estimate the range-dependent interference covariance matrix (ICM). In this paper, we propose an adaptive moving target detection approach for FDA radar without training data, where the ICM is directly estimated by applying a structured generalized likelihood ratio test (SGLRT) to the target-present samples. In order to reduce the three-dimensional (3D) search complexity of the SGLRT, we then propose an efficient unstructured GLRT (UGLRT) approach, which transforms the 3D search into a 1D Doppler-only search and a joint range-angle search. Moreover, we design a low-complexity two-stage implementation of the UGLRT for practical application. The first stage is to test multiple discrete frequency bins for target detection, while the second stage is to precisely estimate the frequency through locally optimizing the estimator function around the coarse estimate provided by the first-stage detection. As the UGLRT approach avoids the 3D search, it significantly reduces the computational complexity, compared to the SGLRT counterpart. Numerical results show that the UGLRT approach achieves almost the same detection and estimation performance as the SGLRT one.

A discrete projection analogue to Pick’s theorem

Pick’s theorem expresses the area of a polygon on a grid in terms of the number of boundary and interior integer lattice points. Here, we present an analogous theorem for the area of a symmetric, convex polygon in terms of the number of polygon edges and total projection bins. These polygons arise naturally through discrete projection ghosts. Ghosts are geometric objects that define locations in discrete tomographic systems which are not uniquely determinable. In this work, we show that the area A of a ghost’s convex hull is related to the number of non-trivial discrete projection bins B over the ghost image for any set of N 2D discrete projections by A=B/2−N/2. The ratio B/A has a strong upper bound of exactly 2. This relation is analogous to Pick’s theorem for polygons with lattice point vertices.

The ALMA Spectroscopic Survey in the HUDF: A Model to Explain Observed 1.1 and 0.85 mm Dust Continuum Number Counts

Abstract We present a new semiempirical model for the dust continuum number counts of galaxies at 1.1 mm and 850 μm. Our approach couples an observationally motivated model for the stellar mass and star formation rate distribution of galaxies with empirical scaling relations to predict the dust continuum flux density of these galaxies. Without a need to tweak the IMF, the model reproduces the currently available observations of the 1.1 mm and 850 μm number counts, including the observed flattening in the 1.1 mm number counts below 0.3 mJy and the number counts in discrete bins of different galaxy properties. Predictions of our work include the following: (1) the galaxies that dominate the number counts at flux densities below 1 mJy (3 mJy) at 1.1 mm (850 μm) have redshifts between z = 1 and z = 2, stellar masses of ∼5 × 1010 M ⊙, and dust masses of ∼108 M ⊙; (2) the flattening in the observed 1.1 mm number counts corresponds to the knee of the 1.1 mm luminosity function. A similar flattening is predicted for the number counts at 850 μm; (3) the model reproduces the redshift distribution of current 1.1 mm detections; and (4) to efficiently detect large numbers of galaxies through their dust continuum, future surveys should scan large areas once reaching a 1.1 mm flux density of 0.1 mJy rather than integrating to fainter fluxes. Our modeling framework also suggests that the amount of information on galaxy physics that can be extracted from the 1.1 mm and 850 μm number counts is almost exhausted.

A continuous binning for discrete, sparse and concentrated observations

Discrete observations from data which are obtained from sparse, and yet concentrated events are often observed (e.g. road accidents or murders). Traditional methods to compute summary statistics often include placing the data in discrete bins but for this type of data this approach often results in large numbers of empty bins for which no function or summary statistic can be computed.Here, a method for dealing with sparse and concentrated observations is constructed, based on a sequence of non-overlapping bins of varying size, which gives a continuous interpolation of data for computing summary statistics of the values for the data, such as the mean.The method presented here overcomes the problem which sparsity and concentration present when computing functions to represent the data. Implementation of the method presented here is facilitated via open access to the code.•A new method for computing functions over sparse and concentrated data is constructed.•The method allows straightforward functions to be computed over partitions of the data, such as the mean, but also more complicated functions, such as coefficients, ratios, correlations, regressions and others.

African Buffalo Optimization for One Dimensional Bin Packing Problem

African Buffalo Optimization (ABO) is one of the most recent bioinspired metaheuristics based on swarm intelligence. It is inspired by the buffalo's behavior and lifestyle. ABO Metaheuristic showed its effectiveness for solving several optimization problems. In this contribution, we present an adaptive ABO for solving the NP-hard one dimensional Bin Packing Problem (1BPP). In the proposed algorithm, we used the ABO algorithm in combination with Ranked Order Value method to obtain discrete values and Bin Packing Problem heuristics to incorporate the problem knowledge. The proposed algorithm is used to solve 1210 of 1BPP instances. The obtained results are compared with those found by recent algorithms in the literature. Computational results show the effectiveness of the proposed algorithm and its ability to achieve best and promising solutions.

Redshift measurement through star formation

In this work we use the property that, on average, star formation rate increases with redshift for objects with the same mass – the so called galaxy main sequence – to measure the redshift of galaxy clusters. We use the fact that the general galaxy population forms both a quenched and a star-forming sequence, and we locate these ridges in the SFR–M⋆ plane with galaxies taken from the Sloan Digital Sky Survey in discrete redshift bins. We fitted the evolution of the galaxy main sequence with redshift using a new method and then subsequently apply our method to a suite of X-ray selected galaxy clusters in an attempt to create a new distance measurement to clusters based on their galaxy main sequence. We demonstrate that although it is possible in several galaxy clusters to measure the main sequences, the derived distance and redshift from our galaxy main sequence fitting technique has an accuracy of σz = ±0.017 ⋅ (z + 1) and is only accurate up to z ≈ 0.2.