143174 research outputs found
Sort by
Redefining the dielectric response of nanoconfined liquids: insights from water
No full textRecent experiments show that the relative dielectric constant of water confined to a film of nanometric
thickness reaches a strikingly low value of 2.1, barely above the bulk’s 1.8 value for the purely electronic
response. We argue that is not a well-defined measure for dielectric properties at subnanometer scales
due to the ambiguous definition of confinement width. Instead, we propose the 2D polarizability α⊥ as the
appropriate, well-defined response function whose magnitude can be directly obtained from both measurements and computations. Once the appropriate description is used, understanding the interplay between electronic and ionic contributions becomes critical, contrary to what is widely assumed. This highlights the importance of electronic degrees of freedom in interpreting the dielectric response of polar fluids under nanoconfinement conditions, as revealed by molecular dynamics simulations
Numerical study on the blast behavior of RC fences strengthened with high-strength strain-hardening cementitious composites
No full textThis paper presents a numerical study on the blast performance of reinforced concrete (RC) fences strengthened with high-strength strain-hardening cementitious composite (HS-SHCC) using finite element (FE) modeling. In the FE model, the MAT_72R3 material model is calibrated and verified against analytical method to accurately simulate the material characteristics of HS-SHCC. A comparative study between RC and HS-SHCC fences subjected to blast loading is conducted to assess the potential of HS-SHCC as a replacement material for protective structures. The results demonstrate that the HS-SHCC fence outperforms the RC fence in terms of smaller maximum and residual deflection, reduced damage, and debris. The study further investigates the strengthening of existing RC fences with layers of HS-SHCC under dynamic and impulsive blast loadings. The response, failure modes, strain distribution, and energy dissipation are compared to identify the most effective strengthening configuration for each blast loading regime. Results show that, for dynamic blast loading, the front strengthened RC fence demonstrates better performance and higher blast resistance due to the high tensile capacity of HS-SHCC, which effectively resists flexural damage. Conversely, the back strengthened RC fence exhibits notable improvement and higher blast resistance under impulsive blast loading as the back HS-SHCC layer protects the RC fence against local punching shear failure
Near-instantaneous atmospheric retrievals and model comparison with FASTER
No full textIn the era of the James Webb Space Telescope (JWST), the dramatic improvement in the spectra of exoplanetary atmospheres demands a corresponding leap forward in our ability to analyze them: atmospheric retrievals need to be performed on thousands of spectra, applying to each large ensembles of models (that explore atmospheric chemistry, thermal profiles, and cloud models) to identify the best one(s). In this limit, traditional Bayesian inference methods such as nested sampling become prohibitively expensive. We introduce Fast Amortized Simulation-based Transiting Exoplanet Retrieval (FASTER), a neural-network-based method for performing atmospheric retrieval and Bayesian model comparison at a fraction of the computational cost of classical techniques. We demonstrate that the marginal posterior distributions of all parameters within a model and the posterior probabilities of the models we consider match those computed using nested sampling both on mock spectra and for the real NIRSpec PRISM spectrum of WASP-39b. The true power of the FASTER framework comes from its amortized nature, which allows the trained networks to perform practically instantaneous Bayesian inference and model comparison over ensembles of spectra—real or simulated—at minimal additional computational cost. This offers valuable insight into the expected results of model comparison (e.g., distinguishing cloudy from cloud-free and isothermal from nonisothermal models), as well as their dependence on the underlying parameters, which is computationally unfeasible with nested sampling. This approach will constitute as large a leap in spectral analysis as the original retrieval methods based on Markov Chain Monte Carlo have proven to be
L1 benchmark definitions and results
No full textThe chapter presents a collection of analytical benchmark problems specifically selected to provide a set of stress tests for the assessment of multi-fidelity optimization methods. These benchmarks, here denoted as L1 problems, address challenges related to the curse of dimensionality and the scalability associated with multi-fidelity methods, handling localized, multimodal, and discontinuous behaviors of the objective functions, and handling the possible presence of noise in the objective functions. The first objective is to provide the community with hard-to-solve but easy-to-implement problems. The second objective is to assess and compare quite a large
variety of multi-fidelity methods against benchmark problems. Multi-fidelity methods performance are assessed in terms of both optimization and global accurac
Memristor-assisted background calibration for analog-to-digital converters (ADCs)
No full textThis thesis describes the main works undertaken to explore the feasibility of exploiting the memristor to improve the linearity of the ADC (i.e., integral non-linearity (INL), differential non-linearity (DNL), signal-to-noise-and-distortion ratio (SNDR), and spurious-free dynamic range (SFDR)). A memristor-assisted background calibration scheme has been proposed. Its functionality and efficacy have been demonstrated in Nyquist successive-approximation-register ADCs (SAR ADCs) with a resistive digital-to-analog converter (RDAC) and a capacitive DAC (CDAC). The complete circuitry is designed in a standard 180 nm and 65 nm Bipolar-CMOS-DMOS (BCD) process. The memristor model used for the circuit design was developed from a physical in-house fabricated memristor and validated in Cadence virtuoso simulation tools. At the end of the research, the proposed calibration scheme is able to calibrate a 100 kS/s 12b SAR ADC’s SNDR from 49.40 dB to 63.70 dB, with a total ADC area of 0.08 mm2 in 180 nm process, and a 5 MS/s 12b SAR ADC’s SNDR from 46.23 dB to 66.10 dB, with a total ADC area of 0.0153 mm2 in 65 nm process. To the best of the author’s knowledge, this is the first research on the memristor-assisted background calibration scheme for ADCs. As a feasibility study, this research is still at the initial stage, which means there are a lot of design challenges and limitations, making this emerging scheme is not yet competitive compared with the state-of-the-art ADCs with conventional CMOS calibration schemes. With the reported results, the author would like to draw attentions of the ADC design community to an alternative calibration scheme enabled by the emerging technology. The author expects the ADC performance will improve in the near future as the emerging CIM device technology (not only the ”memristor”) is advancing rapidly.Open Acces
Multi-fidelity methods
No full textThe multi-fidelity methods assessed by the AVT-331 technical team are described in this chapter. A definition of what constitutes a multi-fidelity method is given, along with a unifying nomenclature. These elements are intended to be as consistent as possible within this chapter and across later chapters to facilitate reading and understanding of the presented results by a broad community. A taxonomy of multi-fidelity methods is proposed whereby nearly 20 multi-fidelity methods studied by the team are categorized into a much smaller set of method classes. The largest class is a set of multi-fidelity surrogate modeling approaches with various online adaptive sampling techniques. Each multi-fidelity method applied by AVT-331 is provided a suitable discussion in this chapter. Not all types of multi-fidelity methods of interest to this group are studied and applied to AVT-331 benchmarks, owing to time and resource constraints. However, a wide range of approaches are evaluated by AVT-331, experience which may be helpful to the NATO vehicle design community
A statically and dynamically scalable soft SIMT processor
No full textThis goal of this project was to create an effective, high performance soft processor. A SIMT architecture aligns well with the parallel nature of the FPGA and also enables comparison with the most successful contemporary parallel processor, the GPU. To be effective, the soft SIMT processor must be small, fast, and efficient. The result of this work is a compact (an area ≈ 1% of a mid-range FPGA), high speed (771 MHz with an unconstrained compile), and performant (similar computational efficiency with contemporary GPUs) core.
The processor is both statically and dynamically scalable. Static scalability is enabled by many different architectural variants and features, which allow the core to use the FPGA resources effectively, and also optimize the processor for different workloads. The dynamic scalability can change the width and depth of the thread space on an instruction-by-instruction basis, which mitigates the main limitation of parallel processors mapped to the FPGA, which is memory bandwidth.
A number of soft GPUs have previously been published, but the large resource requirements and relatively modest performance make them uncompetitive for any real applications. In contrast, the processor in this project is intended to be used for embedded applications in contemporary FPGA designs. In addition to direct comparisons with earlier soft GPUs, the new SIMT core is also benchmarked against commercial GPUs for architectural efficiency and returns a similar effective FLOPs percentage.Open Acces
Evaluating the surface shortwave radiative impact of african landscape fires
No full textWildfires are an important Earth system process with far-reaching impacts. The African continent is responsible for over 70% of all landscape burning worldwide, and over 10% of African land burns every year. The work presented here uses satellite data to investigate the effects of landscape fires in Africa on surface albedo, how these changes affect the surface shortwave radiative balance, and what the associated change in surface temperature is. Surface albedo is found to decrease immediately after a landscape burn due to the charring of the surface, with a continental average decline of 0.019±0.001. Albedo recovers exponentially thereafter, and some level of surface brightening is observed in the long term because of vegetation removal; this is especially prominent in the Kalahari region. Land cover types with higher tree cover experience less albedo change, faster recovery and less brightening after a fire. Landscape fires are found to cause a significant instantaneous surface shortwave radiative forcing (RF), with the average warming effect in fire affected pixels peaking at 4.5±1.7 Wm-2. RF decreases after a fire, following a recovery similar to post-fire albedo. In months 5-10 after a fire, the continent-wide RF is small and negative, driven almost exclusively by behaviour in the Kalahari region. Land surface temperature (LST) increases immediately after a fire, with an average peak LST anomaly of 1.9±0.6 K. This change cannot be explained by the albedo-induced radiative forcing only, suggesting the latent heat flux is also affected by fire. This causes additional heating at the surface, especially in areas with higher tree cover and moderate to high soil moisture. The analysis of 20 years of data reveals few trends in all variables investigated, largely due to high interannual variability of fire and other variables across the continent.Open Acces
Phase space approximations for non-hermitian quantum systems
No full textIn this thesis, we study phase-space approximations to quantum dynamics for both Hermitian and non-Hermitian quantum systems, motivated by the need to efficiently describe quantum effects in models that are analytically intractable but admit classical analogues offering insight into their complex behaviour. Such systems arise in diverse areas, include but are not limited to quantum optics, condensed matter, open quantum systems, and quantum technologies.
We begin by exploring phase-space approximations for quantum systems governed by the Heisenberg-Weyl algebra. Building on established coherent-state methods, we develop a semiclassical Husimi approximation for non-Hermitian Hamiltonians. This approach reveals that the resulting dynamics comprise the transport of initial Husimi distributions along classical trajectories, overlaid with a time-dependent norm landscape reflecting non-unitary evolution. We demonstrate the effectiveness of this method through several examples, including damped Kerr oscillators and -symmetric models, providing a superior approximation for observables compared to standard coherent-state techniques.
To generalise these ideas, we develop a unified framework for phase-space approximations that applies across a wide range of algebraic structures. By identifying a set of general criteria, we enable systematic application of the approximation scheme to any quantum system satisfying these conditions. This provides a cohesive theoretical foundation that connects previously distinct results under a common approach.
We then apply this framework to both the algebra and non-Lie algebraic bosonic conversion systems. For , we analyse two-mode Bose-Hubbard models with various Hamiltonians, demonstrating quantum-classical correspondence and exploring the effects of -symmetry. For bosonic conversion systems, we propose coherent states for these non-Lie algebraic structures and derive their associated phase-space geometry, characterised as quantum Kummer shapes with classical counterparts as orbifolds.
Our results advance the understanding of quantum-classical correspondence in multi-boson systems beyond traditional Hermitian frameworks and offer practical tools for simulating quantum dynamics in experimentally relevant models.Open Acces
An EGR1-dependent cascade modulates genome architecture at the CSF1R locus
No full textThe organization and dynamics of chromatin are key to regulating gene expression during myeloid cell differentiation. Sequence-specific transcription factors initiate and maintain a complex network of enhancer-promoter contacts, which is supported by insulating elements and genome folding organizers such as CTCF and Cohesin. The spatial arrangement of enhancers and promoters, as well as their epigenetic state, drives cell and tissue-specific transcriptomes. Here we dissect the spatial, transcriptional, and epigenetic landscape of the colony stimulating factor 1 receptor (CSF1R) locus in monocytes and macrophages. CSF1R is a receptor tyrosine kinase that triggers the signaling cascade required for macrophage differentiation. Previous work showed that CSF1R expression is regulated by multiple enhancers, including the fms-intronic regulatory element (FIRE). Here, we find that a single EGR-1 binding motif dictates activation of CSF1R. We also discover that the CSF1R entire locus folds into a hub of gene regulation, affecting an extended network of myeloid and inflammatory genes. Globally, EGR1 may have an expanded role as a macrophage-specific boundary element, supporting enhancer-promoter looping at several genes. In sum, we describe a novel 3D chromatin network that is critical for macrophage development and function