Search NASA⌕ Search

SEARCH · Search NASA

Results for “quantum distributed algorithms”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 343 records · Page 19

Astrophysical data analysis with information field theory

Non-parametric imaging and data analysis in astrophysics and cosmology can be addressed by information field theory (IFT), a means of Bayesian, data based inference on spatially distributed signal fields. IFT is a statistical field theory, which permits the construction of optimal signal recovery algorithms. It exploits spatial correlations of the signal fields even for nonlinear and non-Gaussian signal inference problems. The alleviation of a perception threshold for recovering signals of unknown correlation structure by using IFT will be discussed in particular as well as a novel improvement on instrumental self-calibration schemes. IFT can be applied to many areas. Here, applications in in cosmology (cosmic microwave background, large-scale structure) and astrophysics (galactic magnetism, radio interferometry) are presented.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Phase diagram of the hard-core Bose-Hubbard model on a checkerboard superlattice

We obtain the complete phase diagram of the hard-core Bose-Hubbard model in the presence of a period-two superlattice in two and three dimensions. First we acquire the phase boundaries between the superfluid phase and the 'trivial' insulating phases of the model (the completely-empty and completely-filled lattices) analytically. Next, the boundary between the superfluid phase and the half-filled Mott-insulating phase is obtained numerically, using the stochastic series expansion algorithm followed by finite-size scaling. We also compare our numerical results against the predictions of several approximation schemes, including two mean-field approaches and a fourth-order strong-coupling expansion, where we show that the latter method in particular is successful in producing an accurate picture of the phase diagram. Finally, we examine the extent to which several approximation schemes, such as the random phase approximation and the strong-coupling expansion, give an accurate description of the momentum distribution of the bosons inside the insulating phases.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Characterization of Thermal Parameters for Improving Pyranometer and Pyrgeometer Measurements

Since the introduction of thermopile, pyranometers (solar, e.g., 0.3-3.0 micrometers) and pyrgeometers (terrestrial, e.g., 4-50 micrometers) have become instruments commonly used for measuring the broadband hemispherical irradiances at the surface in a long-term, monitoring mode for decades. These commercially available radiometers have been manufactured in several countries such as from the United States, Asia, and Europe, and are generally reliable and economical. These worldwide distributions of surface measurements become even more important in the era of Earth remote sensing in studying climate change. However, recent studies from field campaigns have pointed out that erroneous factors (e.g., temperature gradients between the filter dome and detector, emissivity of the thermopile) are responsible for the unacceptable level of uncertainty (e.g., 20 W m(exp -2)). Using a newly developed instrument of Quantum Well Infrared Photodetector (QWTP), we have characterized the brightness temperature fields of pyranometers and pyrgeometers under various sky conditions. The QWIP is based on the superlattice (GaAs/AlGaAs) technology and has a noise equivalent temperature (NEAT) less than 0.1 K. The quality of pyranometer and pyrgeometer measure- ments can be improved largely by applying proper knowledge of the thermal parameters affecting the operation of the thermopile systems. Data correction procedure and algorithm will be presented and discussed.

Tsay, Si-Chee↗

Mesoscale Simulations of Particulate Flows with Parallel Distributed Lagrange Multiplier Technique

Fluid particulate flows are common phenomena in nature and industry. Modeling of such flows at micro and macro levels as well establishing relationships between these approaches are needed to understand properties of the particulate matter. We propose a computational technique based on the direct numerical simulation of the particulate flows. The numerical method is based on the distributed Lagrange multiplier technique following the ideas of Glowinski et al. (1999). Each particle is explicitly resolved on an Eulerian grid as a separate domain, using solid volume fractions. The fluid equations are solved through the entire computational domain, however, Lagrange multiplier constrains are applied inside the particle domain such that the fluid within any volume associated with a solid particle moves as an incompressible rigid body. Mutual forces for the fluid-particle interactions are internal to the system. Particles interact with the fluid via fluid dynamic equations, resulting in implicit fluid-rigid-body coupling relations that produce realistic fluid flow around the particles (i.e., no-slip boundary conditions). The particle-particle interactions are implemented using explicit force-displacement interactions for frictional inelastic particles similar to the DEM method of Cundall et al. (1979) with some modifications using a volume of an overlapping region as an input to the contact forces. The method is flexible enough to handle arbitrary particle shapes and size distributions. A parallel implementation of the method is based on the SAMRAI (Structured Adaptive Mesh Refinement Application Infrastructure) library, which allows handling of large amounts of rigid particles and enables local grid refinement. Accuracy and convergence of the presented method has been tested against known solutions for a falling sphere as well as by examining fluid flows through stationary particle beds (periodic and cubic packing). To evaluate code performance and validate particle contact physics algorithm, we performed simulations of a representative experiment conducted at the University of California at Berkley for pebble flow through a narrow opening.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Investigation of an apodized imaged Hartmann wavefront sensor

Here, quantitative wavefront measurements are demonstrated using a Hartmann mask re-imaged onto a camera. The wavefront is reconstructed using standard algorithms applied to the difference of beamlet centroids determined from fluence distributions obtained for two different longitudinal locations of the mask. The wavefront of the optical wave in the object plane is measured independently of imaging-system collimation. Apodization obtained with spatially dithered distributions of small transparent or opaque pixels improves the measurement accuracy by reducing the spatial-frequency content of the mask holes. Simulations and experiments demonstrate the excellent accuracy of this diagnostic over a wide range of parameters, making it suitable, for example, to characterize laser systems.

47 OTHER INSTRUMENTATION↗

A combined Event-Driven/Time-Driven molecular dynamics algorithm for the simulation of shock waves in rarefied gases

A novel combined Event-Driven/Time-Driven (ED/TD) algorithm to speed-up the Molecular Dynamics simulation of rarefied gases using realistic spherically symmetric soft potentials is presented. Due to the low density regime, the proposed method correctly identifies the time that must elapse before the next interaction occurs, similarly to Event-Driven Molecular Dynamics. However, each interaction is treated using Time-Driven Molecular Dynamics, thereby integrating Newton's Second Law using the sufficiently small time step needed to correctly resolve the atomic motion. Although infrequent, many-body interactions are also accounted for with a small approximation. The combined ED/TD method is shown to correctly reproduce translational relaxation in argon, described using the Lennard-Jones potential. For densities between {rho}=10{sup -4}kg/m{sup 3} and {rho}=10{sup -1}kg/m{sup 3}, comparisons with kinetic theory, Direct Simulation Monte Carlo, and pure Time-Driven Molecular Dynamics demonstrate that the ED/TD algorithm correctly reproduces the proper collision rates and the evolution toward thermal equilibrium. Finally, the combined ED/TD algorithm is applied to the simulation of a Mach 9 shock wave in rarefied argon. Density and temperature profiles as well as molecular velocity distributions accurately match DSMC results, and the shock thickness is within the experimental uncertainty. For the problems considered, the ED/TD algorithm ranged from several hundred to several thousand times faster than conventional Time-Driven MD. Moreover, the force calculation to integrate the molecular trajectories is found to contribute a negligible amount to the overall ED/TD simulation time. Therefore, this method could pave the way for the application of much more refined and expensive interatomic potentials, either classical or first-principles, to Molecular Dynamics simulations of shock waves in rarefied gases, involving vibrational nonequilibrium and chemical reactivity.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Advances in machine-learning-based sampling motivated by lattice quantum chromodynamics

Sampling from known probability distributions is a ubiquitous task in computational science, underlying calculations in domains from linguistics to biology and physics. Generative machine-learning (ML) models have emerged as a promising tool in this space, building on the success of this approach in applications such as image, text, and audio generation. Often, however, generative tasks in scientific domains have unique structures and features—such as complex symmetries and the requirement of exactness guarantees—that present both challenges and opportunities for ML. This Perspective outlines the advances in ML-based sampling motivated by lattice quantum field theory, in particular for the theory of quantum chromodynamics. Enabling calculations of the structure and interactions of matter from our most fundamental understanding of particle physics, lattice quantum chromodynamics is one of the main consumers of open-science supercomputing worldwide. Here, the design of ML algorithms for this application faces profound challenges, including the necessity of scaling custom ML architectures to the largest supercomputers, but also promises immense benefits, and is spurring a wave of development in ML-based sampling more broadly. In lattice field theory, if this approach can realize its early promise it will be a transformative step towards first-principles physics calculations in particle, nuclear and condensed matter physics that are intractable with traditional approaches.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A mathematical model of the passage of an asteroid-comet body through the Earth’s atmosphere

In the paper, a mathematical model and a numerical algorithm are proposed for modeling the complex of phenomena which accompany the passage of a friable asteroid-comet body through the Earth’s atmosphere: the material ablation, the dissociation of molecules, and the radiation. The proposed model is constructed on the basis of the Navier-Stokes equations for viscous heat-conducting gas with an additional equation for the motion and propagation of a friable lumpy-dust material in air. The energy equation is modified for the relation between two its kinds: the usual energy of the translation of molecules (which defines the temperature and pressure) and the combined energy of their rotation, oscillation, electronic excitation, dissociation, and radiation. For the mathematical model of atmosphere, the distribution of density, pressure, and temperature in height is taken as for the standard atmosphere. An asteroid-comet body is taken initially as a round body consisting of a friable lumpy-dust material with corresponding density and significant viscosity which far exceed those for the atmosphere gas. A numerical algorithm is proposed for solving the initial-boundary problem for the extended system of Navier-Stokes equations. The algorithm is the combination of the semi-Lagrangian approximation for Lagrange transport derivatives and the conforming finite element method for other terms. The implementation of these approaches is illustrated by a numerical example.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Scaling whole-chip QAOA for higher-order ising spin glass models on heavy-hex graphs

Abstract We show that the quantum approximate optimization algorithm (QAOA) for higher-order, random coefficient, heavy-hex compatible spin glass Ising models has strong parameter concentration across problem sizes from 16 up to 127 qubits for p = 1 up to p = 5, which allows for computationally efficient parameter transfer of QAOA angles. Matrix product state (MPS) simulation is used to compute noise-free QAOA performance. Hardware-compatible short-depth QAOA circuits are executed on ensembles of 100 higher-order Ising models on noisy IBM quantum superconducting processors with 16, 27, and 127 qubits using QAOA angles learned from a single 16-qubit instance using the JuliQAOA tool. We show that the best quantum processors find lower energy solutions up to p = 2 or p = 3, and find mean energies that are about a factor of two off from the noise-free distribution. We show that p = 1 QAOA energy landscapes remain very similar as the problem size increases using NISQ hardware gridsearches with up to a 414 qubit processor.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Cross-scale Efficient Tensor Contractions for Coupled Cluster Computations Through Multiple Programming Model Backends

Coupled-cluster methods provide highly accurate models of molecular structure by explicit numerical calculation of tensors representing the correlation between electrons. These calculations are dominated by a sequence of tensor contractions, motivating the development of numerical libraries for such operations. While based on matrix-matrix multiplication, these libraries are specialized to exploit symmetries in the molecular structure and in electronic interactions, and thus reduce the size of the tensor representation and the complexity of contractions. The resulting algorithms are irregular and their parallelization has been previously achieved via the use of dynamic scheduling or specialized data decompositions. We introduce our efforts to extend the Libtensor framework to work in the distributed memory environment in a scalable and energy efficient manner. We achieve up to 240 speedup compared with the best optimized shared memory implementation. We attain scalability to hundreds of thousands of compute cores on three distributed-memory architectures, (Cray XC30&XC40, BlueGene/Q), and on a heterogeneous GPU-CPU system (Cray XK7). As the bottlenecks shift from being compute-bound DGEMM's to communication-bound collectives as the size of the molecular system scales, we adopt two radically different parallelization approaches for handling load-imbalance. Nevertheless, we preserve a uni ed interface to both programming models to maintain the productivity of computational quantum chemists.

74 ATOMIC AND MOLECULAR PHYSICS↗

Analysis of Three Body Decays in Quasi-Real Photoproduction

This thesis presents preliminary measurements of the moments of angular distribution for the K?K+ mesonic final state using a quasi-real, linearly polarised photon. Two of the main points of focus for this analysis were the development of the formalism for photoproduced vector-pseudoscalar decay and the tools by which the analysis was carried out. The formalism described how the moments of angular distribution could be extracted from the angular decay variables, and how they related to resonance decays in terms of partial waves. In extracting the moments, it was found that adapted Markov Chain Monte Carlo (MCMC) methods proved to be more effective in terms of extracting results from the data compared to the sole use of gradient descent based fitting algorithms. These measurements are complementary to other decay channels and production mechanisms, with the global aim of probing the existence of exotic mesons. Data from this analysis was taken by the CLAS collaboration at the Thomas Jefferson National Accelerator Facility (JLab) by scattering from a liquid hydrogen target using a 10.6 GeV electron. This work was done within the MesonEx program at CLAS12 that aims to map the spectrum of mesons, and in doing so, gain a greater understanding of Quantum ChromoDynamics (QCD) and the strong force.

Wishart, Robert↗

SU-F-T-145: Modeling of a Proton Pencil Beam Partially Blocked by a Collimator

Purpose: To develop a mathematical model of the fluence distribution of a proton pencil beam that is partially blocked by a collimator. Methods: In one dimension, an exact analytical solution of the proton distribution behind a collimator edge was derived, taking the beam’s phase space, the edge position, and the scattering downstream of the collimator as input. Extension to two dimensions was achieved using a Clarkson-like Method: determine the radial distances from each calculation point to the collimator outline, calculate the beam’s phase space in the direction of each radial line, apply the one-dimensional model along each radial line, and sum the contributions. A Matlab implementation of the full model produced high-resolution distributions with relatively short calculation times (of about 2 ms per calculation point). The model was validated by comparison to Monte Carlo simulations for select pencil beam phase space distributions and collimator shapes. Results: Agreement between model and simulation confirm the exact mathematical solution of the one-dimensional model. In two dimensions the agreement is reasonable, but not exact, due to the approximations introduced in the Clarkson-like method. Conclusion: A model has been developed and validated that gives the lateral fluence distribution of a pencil beam beyond a collimator. Notably, this model does not include free parameters and the final fluence is completely determined by the phase space parameters of the beam, already existing in the treatment planning system, as well as both axial and lateral position of the aperture edge. As such, it can be incorporated in a full pencil-beam dose calculation algorithm allowing for the modeling of field-specific collimators in proton pencil beam scanning and, potentially, the optimization of collimator shape and/or pencil phase space in addition to the customary spot weight optimization.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Coarse-Grained Density Functional Theory Predictions via Deep Kernel Learning

Scalable electronic predictions are critical for soft materials design. Recently, the Electronic Coarse-Graining (ECG) method was introduced to renormalize all-atom quantum chemical (QC) predictions to coarse-grained (CG) resolutions using deep neural networks (DNNs). While DNNs can learn complex representations that prove challenging for kernel-based methods, they are susceptible to overfitting and the overconfidence of uncertainty estimations. Here, we develop ECG within a GPU-accelerated Deep Kernel Learning (DKL) framework to enable CG QC predictions using range-separated hybrid density functional theory (DFT), obtaining a 107 speedup relative to naive all-atom QC. By treating the predicted electronic properties as random Gaussian Processes, DKL incorporates CG mapping degeneracy by learning the distribution of electronic energies as a function of CG configuration. DKL-ECG accurately reproduces molecular orbital energies from range-separated DFT while facilitating efficient training via active learning using the uncertainties provided by DKL. Further, we show that while active learning algorithms enable efficient sampling of a more diverse configurational space relative to random sampling, all explored query methods exhibit comparable performance for the examined system. We attribute this result to the significant overlap of the feature space and output property distributions across multiple temperatures.

97 MATHEMATICS AND COMPUTING↗

Spinor representations for fields with any spin: Lorentz tensor basis for operators and covariant multipole decomposition

This paper discusses a framework to parametrize and decompose operator matrix elements for particles with higher spin (j > 1/2) using chiral representations of the Lorentz group, i.e. the (j, 0) and (0, j) representations and their parity-invariant direct sum. Unlike traditional approaches that require imposing constraints to eliminate spurious degrees of freedom, these chiral representations contain exactly the 2j + 1 components needed to describe a spin-j particle. The central objects in the construction are the t-tensors, which are generalizations of the Pauli four-vector σ μ for higher spin. For the generalized spinors of these representations, we demonstrate how the algebra of the t-tensors allows to formulate a generalization of the Dirac matrix basis for any spin. For on-shell bilinears, we show that a set consisting exclusively of covariant multipoles of order 0 ≤ m ≤ 2j forms a complete basis. We provide explicit expressions for all bilinears of the generalized Dirac matrix basis, which are valid for any spin value. As a byproduct of our derivations we present an efficient algorithm to compute the t-tensor matrix elements. The formalism presented here paves the way to use a more unified approach to analyze the non-perturbative QCD structure of hadrons and nuclei across different spin values, with clear physical interpretation of the resulting distributions as covariant multipoles.

Angular momentum of light↗

ThickBrick: optimal event selection and categorization in high energy physics. Part I. Signal discovery

We provide a prescription called ThickBrick to train optimal machine-learning-based event selectors and categorizers that maximize the statistical significance of a potential signal excess in high energy physics (HEP) experiments, as quantified by any of six different performance measures. For analyses where the signal search is performed in the distribution of some event variables, our prescription ensures that only the information complementary to those event variables is used in event selection and categorization. This eliminates a major misalignment with the physics goals of the analysis (maximizing the significance of an excess) that exists in the training of typical ML-based event selectors and categorizers. In addition, this decorrelation of event selectors from the relevant event variables prevents the background distribution from becoming peaked in the signal region as a result of event selection, thereby ameliorating the challenges imposed on signal searches by systematic uncertainties. Our event selectors (categorizers) use the output of machine-learning-based classifiers as input and apply optimal selection cutoffs (categorization thresholds) that are functions of the event variables being analyzed, as opposed to flat cutoffs (thresholds). These optimal cutoffs and thresholds are learned iteratively, using a novel approach with connections to Lloyd’s k-means clustering algorithm. We provide a public, Python implementation of our prescription, also called ThickBrick, along with usage examples.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Turbulent flow characteristics in an 84-pin rod bundle for typical and damaged spacer grids

Hexagonal rod bundles arranged in a tightly packed triangular lattice are extensively used for heat transfer and energy generation applications. Staggered spacer grids are used to maintain the structural integrity of gas-cooled fast reactor (GFR) fuel assemblies, while inducing localized turbulence in flow. Damage to these spacer grids results in a disruption of flow fields within these hexagonal fuel bundles. Experimental flow visualizations are critical to identify the differences in local flow properties that the structural damage may cause. This experimental research investigates the flow-field characteristics at a near-wall and center plane in a prototypical 84-pin GFR fuel assembly. Newly installed typical spacers and spacers subject to naturally occurring damage due to material degradation over prolonged experimentation were investigated. Velocity fields were acquired by utilizing the matched-index-of-refraction method to obtain time-resolved particle image velocimetry measurements for a Reynolds number of 12 000. Reynolds decomposition statistical results divulged differences in the time-averaged velocity, velocity fluctuations, flow anisotropy, and Reynolds stress distributions. Galilean decomposition demarcated the influence of spacer grid damage on the velocity fields. To extract turbulent structures and elucidate mechanisms of flow instabilities, proper orthogonal decomposition analysis was employed. Reduced order flow reconstructions enabled the application of vortex identification algorithms to determine the spatial and statistical characteristics of vortices generated. This research work provides unique experimental data on the spacer grid condition-dependent flow. The results offer a deeper understanding of fluid dynamics behavior to support GFR rod bundle design efforts and computational fluid dynamics model validation.

36 MATERIALS SCIENCE↗

Post-transition state dynamics and product energy partitioning following thermal excitation of the F∙∙∙HCH 2 CN transition state: Disagreement with experiment

Born-Oppenheimer direct dynamics simulations were performed to study atomistic details of the F + CH 3 CN → HF + CH 2 CN H-atom abstraction reaction. The simulation trajectories were calculated with a combined M06-2X/MP2 algorithm utilizing the 6-311++G** basis set. In accord with experiment and assuming the accuracy of transition state theory (TST), the trajectories were initiated at the F-HCH 2 CN abstraction TS with a 300 K Boltzmann distribution of energy and directed towards products. Recrossing of the TS was negligible, confirming the accuracy of TST for the simulation. HF formation was rapid, occurring within 0.014 ps of the trajectory initiation. The intrinsic reaction coordinate (IRC) for reaction involves rotation of HF about CH 2 CN and then trapping in the CH 2 CN-HF post-reaction potential energy well of ~10 kcal/mol with respect to the HF + CH 2 CN products. In contrast to this IRC, five different trajectory types were observed, with the majority involving direct dissociation and only 11% approximately following the IRC. The HF vibrational and rotational quantum numbers, n and J, were calculated when HF was initially formed and they increase as potential energy is released in forming the HF + CH 2 CN products. The population of the HF product vibrational states is only in qualitative agreement with experiment, with the simulations showing depressed and enhanced populations of the n = 1 and 2 states as compared to experiment. From the simulations and with an anharmonic zero-point energy constraint, the percentage partitioning of the product energy to relative translation, HF rotation, HF vibration, CH 2 CN rotation and CH 2 CN vibration is 5, 11, 60, 7, and 16%, respectively. In contrast the experimental energy partitioning percentages to HF rotation and vibration are 6 and 41%. Comparisons are made between the current simulation and those for other F + H-atom abstraction reactions. The simulation product energy partitioning and HF vibrational population for F + CH 3 CN → HF + CH 2 CN are similar to those for these other reactions. Finally, a detailed discussion is given of possible origins of the difference between the simulation and experimental energy partitioning dynamics for the F + CH 3 CN → HF + CH 2 CN reaction. The F + CH 3 CN reaction also forms the CH 3 C(F)N intermediate, in which the F-atom adds to the C≡N bond. However, this intermediate and the F---CH 3 CN and CH 3 CN-F van der Waals complexes are not expected to affect the F + CH 3 CN → HF + CH 2 CN product energy partitioning.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Preliminary results of teleseismic double-difference relocation of earthquakes around Indonesian archipelago region

Indonesian archipelago region is located in active tectonic setting and high seismicity zone. During the last decade, Indonesian was experienced with destructive major earthquakes causing damage and victims. The information of precise earthquake location parameters are very important in partular for earthquake early warning to the society and for advance seismic studies. In this study, we attempted to improve hypocenter location compiled by BMKG for time periods of April, 2009 up to June, 2014 for about 22,000 earthquake events around Indonesian region. For the firts time, we applied teleseismic double-difference relocation algorithm (teletomoDD) to improve hypocenter region in Indonesia region combining regional and teleseismic stations. Hypocenter relocation was performed utilizing local, regional, and teleseismic P-wave arrival time data. Our relocation result show that travel-time RMS errors were greatly reduced compared to the BMKG catalog. Seismicity at shallower depth (less than 50 km) shows significantly improvement especially in depth, and refined shallow geological structures, e.g. trench and major strike slip faults. Clustered seismicity is also detected beneath volcanic region, and probably related volcano activities and also major faults nearby. In the Sunda arc region, seismicity at shallower depth centered at two major distributions parallel to the trench strike direction, i.e. around fore-arc and in mainland that related to major fault, e.g. the Sumatran fault, and volcanic fronts. Below Central Java region, relocated hypocenter result showed double seismic zone pattern. A seismic gap is detected around the Sunda-Banda transition zone where transition between oceanic subduction to continental crust collision of Australian plate occurs. In Eastern Indonesia region, shallow earthquakes are observed related to major strike slip faults, e.g. Sorong and Palu-Koro fault, volcanism, and shallow part of subduction and collision zones. We also compare our result in the Sunda Arc region with slab1.0 model and our relocated seismicity shows good agreement with the previous slab geometry. Horizontal position shift of relocated events are mostly perpendicular to the trench directions.

58 GEOSCIENCES↗