Search NASA⌕ Search

SEARCH · Search NASA

Results for “generalization bounds”

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 91 records · Page 5

Perturbation Bounds for Markov Chains with General State Space

The aim of this paper is to investigate the stability of Markov chains with general state space. We present new conditions for the strong stability of Markov chains after a small perturbation of their transition kernels. Also, we obtain perturbation bounds with respect to different quantities.

97 MATHEMATICS AND COMPUTING↗

Proof of a quantum Bousso bound

In this work, we prove the generalized Covariant Entropy Bound, $\Delta S\leq (A-A')/4G\hbar$, for light-sheets with initial area $A$ and final area $A'$. The entropy $\Delta S$ is defined as a difference of von Neumann entropies of an arbitrary state and the vacuum, with both states restricted to the light-sheet under consideration. The proof applies to free fields, in the limit where gravitational backreaction is small. We do not assume the null energy condition. In regions where it is violated, we postulate that the bound is protected by the defining property of light-sheets: that their null generators are nowhere expanding.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Interaction between two nonthreshold bound states

A general nonthreshold Bogomol'nyi-Prasad-Sommerfield (F,D{sub p})[or (D{sub p-2},D{sub p})] bound state can be described by a boundary state with a quantized world-volume electric (or magnetic) flux and is characterized by a pair of integers (m,n). With this, we calculate explicitly the interaction amplitude between two such nonthreshold bound states with a separation Y when each of the states is characterized by a pair of integers (m{sub i},n{sub i}) with i=1,2. With this result, one can show that the nondegenerate (i.e., m{sub i}n{sub i}{ne}0) interaction is, in general, attractive for the case of (D{sub p-2},D{sub p}), but this is true and for certain only at large separation for the case of (F,D{sub p}). In either case, this interaction vanishes only if m{sub 1}/n{sub 1}=m{sub 2}/n{sub 2} and n{sub 1}n{sub 2}>0. We also study the analytic structure of the corresponding amplitude and calculate, in particular, the rate of pair production of open strings in the case of (F,D{sub p})

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Bootstrapping the O(N ) vector models

We study the conformal bootstrap for 3D CFTs with O(N) global symmetry. We obtain rigorous upper bounds on the scaling dimensions of the first O(N) singlet and symmetric tensor operators appearing in the $\phi_i \times \phi_j$ OPE, where $\phi_i$ is a fundamental of O(N). Comparing these bounds to previous determinations of critical exponents in the O(N) vector models, we find strong numerical evidence that the O(N) vector models saturate the bootstrap constraints at all values of N. We also compute general lower bounds on the central charge, giving numerical predictions for the values realized in the O(N) vector models. We compare our predictions to previous computations in the 1/N expansion, finding precise agreement at large values of N.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Planck 2013 results. XXIV. Constraints on primordial non-Gaussianity

The Planck nominal mission cosmic microwave background (CMB) maps yield unprecedented constraints on primordial non-Gaussianity (NG). Using three optimal bispectrum estimators, separable template-fitting (KSW), binned, and modal, we obtain consistent values for the primordial local, equilateral, and orthogonal bispectrum amplitudes, quoting as our final result f NL local = 2.7 ± 5.8, f NL equil = -42 ± 75, and f NL orth = -25 ± 39 (68% CL statistical). Non-Gaussianity is detected in the data; using skew-C &ell; statistics we find a nonzero bispectrum from residual point sources, and the integrated-Sachs-Wolfe-lensing bispectrum at a level expected in the ΛCDM scenario. The results are based on comprehensive cross-validation of these estimators on Gaussian and non-Gaussian simulations, are stable across component separation techniques, pass an extensive suite of tests, and are confirmed by skew-C &ell; , wavelet bispectrum and Minkowski functional estimators. Beyond estimates of individual shape amplitudes, we present model-independent, three-dimensional reconstructions of the Planck CMB bispectrum and thus derive constraints on early-Universe scenarios that generate primordial NG, including general single-field models of inflation, excited initial states (non-Bunch-Davies vacua), and directionally-dependent vector models. We provide an initial survey of scale-dependent feature and resonance models. These results bound both general single-field and multi-field model parameter ranges, such as the speed of sound, c s ≥ 0.02 (95% CL), in an effective field theory parametrization, and the curvaton decay fraction r D ≥ 0.15 (95% CL). The Planck data significantly limit the viable parameter space of the ekpyrotic/cyclic scenarios. The amplitude of the four-point function in the local model τ NL < 2800 (95% CL). Taken together, these constraints represent the highest precision tests to date of physical mechanisms for the origin of cosmic structure.

79 ASTRONOMY AND ASTROPHYSICS↗

Limits to the Optical Response of Graphene and Two-Dimensional Materials

Two-dimensional (2D) materials provide a platform for strong light–matter interactions, creating wide-ranging design opportunities via new-material discoveries and new methods for geometrical structuring. Here we derive general upper bounds to the strength of such light–matter interactions, given only the optical conductivity of the material, including spatial nonlocality, and otherwise independent of shape and configuration. Our material figure-of-merit shows that highly doped graphene is an optimal material at infrared frequencies, whereas single-atomic-layer silver is optimal in the visible. For quantities ranging from absorption and scattering to near-field spontaneous-emission enhancements and radiative heat transfer, we consider canonical geometrical structures and show that in certain cases the bounds can be approached, while in others there may be significant opportunity for design improvement. The bounds can encourage systematic improvements in the design of ultrathin broadband absorbers, 2D antennas, and near-field energy harvesters.

2D materials↗

Generalized parton distributions of nuclei

We review recent theoretical results on generalized parton distributions (GPDs) of nuclei, emphasizing the following three roles of nuclear GPDs: (i) complementarity to free proton GPDs, (ii) the enhancement of traditional nuclear effects such nuclear binding, EMC effect, nuclear shadowing, and (iii) an access to novel nuclear effects such as medium modifications of bound nucleons.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Correlated quadratures of resonance fluorescence and the generalized uncertainty relation

Resonance fluorescence from a two-state atom has been predicted to exhibit quadrature squeezing below the Heisenberg uncertainty limit, provided that the optical parameters (Rabi frequency, detuning, laser linewidth, etc.) are chosen carefully. When the correlation between two quadratures of the radiation field does not vanish, however, the Heisenberg limit for quantum fluctuations might be an unrealistic lower bound. A generalized uncertainty relation, due to Schroedinger, takes into account the possible correlation between the quadrature components of the radiation, and it suggests a modified definition of squeezing. We show that the coherence between the two levels of a laser-driven atom is responsible for the correlation between the quadrature components of the emitted fluorescence, and that the Schrodinger uncertainty limit increases monotonically with the coherence. On the other hand, the fluctuations in the quadrature field diminish with an increasing coherence, and can disappear completely when the coherence reaches 1/2, provided that certain phase relations hold.

Arnoldus, Henk F.↗

Scenario Grouping and Decomposition Algorithms for Chance-Constrained Programs

A lower bound for a finite-scenario-based chance-constrained program is the quantile value corresponding to the sorted optimal objective values of scenario subproblems. This quantile bound can be improved by grouping subsets of scenarios at the expense of solving larger subproblems. The quality of the bound depends on how the scenarios are grouped. In this paper, we formulate a mixed-integer bilevel program that optimally groups scenarios to tighten the quantile bounds. For general chance-constrained programs, we propose a branch-and-cut algorithm to optimize the bilevel program, and for chance-constrained linear programs, a mixed-integer linear-programming reformulation is derived. Here, we also propose several heuristics for grouping similar or dissimilar scenarios. Our computational results demonstrate that optimal grouping bounds are much tighter than heuristic bounds, resulting in smaller root-node gaps and better performance of scenario decomposition for solving chance-constrained 0-1 programs. Also, the optimal grouping bounds can be greatly strengthened using larger group size.

97 MATHEMATICS AND COMPUTING↗

Entanglement purification with double selection

We investigate an entanglement purification protocol with double-selection process, which works under imperfect local operations. Compared with the usual protocol with single selection, this double-selection method has higher noise thresholds for the local operations and quantum communication channels and achieves higher fidelity of purified states. It also provides a yield comparable to that of the usual protocol with single selection. We discuss on general grounds how some of the errors which are introduced by local operations are left as intrinsically undetectable. The undetectable errors place a general upper bound on the purification fidelity. The double selection is a simple method to remove all the detectable errors in the first order, so that the upper bound on the fidelity is achieved in the low-noise regime. The double selection is further applied to purification of multipartite entanglement such as two-colorable graph states.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Entropy production fluctuations encode collective behavior in active matter

We derive a general lower bound on distributions of entropy production in interacting active matter systems. The bound is tight in the limit that interparticle correlations are small and short-ranged, which we explore in four canonical active matter models. In all models studied, the bound is weak where collective fluctuations result in long-ranged correlations, which subsequently links the locations of phase transitions to enhanced entropy production fluctuations. We develop a theory for the onset of enhanced fluctuations and relate it to specific phase transitions in active Brownian particles. Furthermore, we also derive optimal control forces that realize the dynamics necessary to tune dissipation and manipulate the system between phases. In so doing, we uncover a general relationship between entropy production and pattern formation in active matter, as well as ways of controlling it.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Net diffusivity in ocean general circulation models with nonuniform grids

The numerical vertical diffusivity K(num), embedded in a numerical ocean general circulation model with nonuniform vertical grid, is estimated. It is shown that in a downwelling region, K(num) is negative for a grid with grid size increasing with depth. When the grid size increment, or the downward vertical velocity, is large, K(num) may exceed the vertical diffusivity specified and may result in a negative effective vertical diffusivity. Therefore care needs to be taken to specify the vertical diffusivity in a numerical model with nonuniform grid, and a lower bound is generally imposed in order to avoid an unphysical negative value. Some possible effects of the negative effective diffusivity are discussed.

Yin, F. L.↗

Large-eddy simulations with wall models

The near-wall viscous and buffer regions of wall-bounded flows generally require a large expenditure of computational resources to be resolved adequately, even in large-eddy simulation (LES). Often as much as 50% of the grid points in a computational domain are devoted to these regions. The dense grids that this implies also generally require small time steps for numerical stability and/or accuracy. It is commonly assumed that the inner wall layers are near equilibrium, so that the standard logarithmic law can be applied as the boundary condition for the wall stress well away from the wall, for example, in the logarithmic region, obviating the need to expend large amounts of grid points and computational time in this region. This approach is commonly employed in LES of planetary boundary layers, and it has also been used for some simple engineering flows. In order to calculate accurately a wall-bounded flow with coarse wall resolution, one requires the wall stress as a boundary condition. The goal of this work is to determine the extent to which equilibrium and boundary layer assumptions are valid in the near-wall regions, to develop models for the inner layer based on such assumptions, and to test these modeling ideas in some relatively simple flows with different pressure gradients, such as channel flow and flow over a backward-facing step. Ultimately, models that perform adequately in these situations will be applied to more complex flow configurations, such as an airfoil.

Cabot, W.↗

Circles-in-the-sky searches and observable cosmic topology in a flat universe

In a universe with a detectable nontrivial spatial topology, the last scattering surface contains pairs of matching circles with the same distribution of temperature fluctuations--the so-called circles-in-the-sky. Searches for nearly antipodal circles-in-the-sky in maps of cosmic microwave background radiation have so far been unsuccessful. This negative outcome, along with recent theoretical results concerning the detectability of nearly flat compact topologies, is sufficient to exclude a detectable nontrivial topology for most observers in very nearly flat positively and negatively curved universes, whose total matter-energy density satisfies 0<|{Omega}{sub tot}-1| < or approx. 10{sup -5}. Here, we investigate the consequences of these searches for observable nontrivial topologies if the Universe turns out to be exactly flat ({Omega}{sub tot}=1). We demonstrate that in this case, the conclusions deduced from such searches can be radically different. We show that, although there is no characteristic topological scale in the flat manifolds, for all multiply-connected orientable flat manifolds, it is possible to directly study the action of the holonomies in order to obtain a general upper bound on the angle that characterizes the deviation from antipodicity of pairs of matching circles associated with the shortest closed geodesic. This bound is valid for all observers and all possible values of the compactification length parameters. We also show that in a flat universe, there are observers for whom the circles-in-the-sky searches already undertaken are insufficient to exclude the possibility of a detectable nontrivial spatial topology. It is remarkable how such small variations in the spatial curvature of the Universe, which are effectively indistinguishable geometrically, can have such a drastic effect on the detectability of cosmic topology. Another important outcome of our results is that they offer a framework with which to make statistical inferences from future circles-in-the-sky searches on whether the Universe is exactly flat.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Coefficient-to-Basis Network: a fine-tunable operator learning framework for inverse problems with adaptive discretizations and theoretical guarantees

We propose a Coefficient-to-Basis Network (C2BNet), a novel framework for solving inverse problems within the operator learning paradigm. C2BNet efficiently adapts to different discretizations through fine-tuning, using a pre-trained model to significantly reduce computational cost while maintaining high accuracy. Unlike traditional approaches that require retraining from scratch for new discretizations, our method enables seamless adaptation without sacrificing predictive performance. Furthermore, we establish theoretical approximation and generalization error bounds for C2BNet by exploiting low-dimensional structures in the underlying datasets. Our analysis demonstrates that C2BNet adapts to low-dimensional structures without relying on explicit encoding mechanisms, highlighting its robustness and efficiency. To validate our theoretical findings, we conducted extensive numerical experiments that showcase the superior performance of C2BNet on several inverse problems. The results confirm that C2BNet effectively balances computational efficiency and accuracy, making it a promising tool to solve inverse problems in scientific computing and engineering applications.

97 MATHEMATICS AND COMPUTING↗

Learning concave-convex profiles of data transport over dedicated connections

Dedicated data transport infrastructures are increasingly being deployed to support distributed big-data and high-performance computing scenarios. These infrastructures employ data transfer nodes that use sophisticated software stacks to support network transport among sites, which often house distributed file and storage systems. Throughput measurements collected over such infrastructures for a range of round trip times (RTTs) reflect the underlying complex end-to-end connections, and have revealed dichotomous throughput profiles as functions of RTT. In particular, concave regions of throughput profiles at lower RTTs indicate near-optimal performance, and convex regions at higher RTTs indicate bottlenecks due to factors such as buffer or credit limits. We present a machine learning method that explicitly infers these concave and convex regions and transitions between them using sigmoid functions. We also provide distribution-free confidence estimates for the generalization error of these concave-convex profile estimates. Throughput profiles for data transfers over 10 Gbps connections with 0–366 ms RTT provide important performance insights, including the near optimality of transfers performed with the XDD tool between XFS filesystems, and the performance limits of wide-area Lustre extensions using LNet routers. A direct application of generic machine learning packages does not adequately highlight these critical performance regions or provide as precise confidence estimates.

Rao, Nageswara↗

Machine Learning Methods for Connection RTT and Loss Rate Estimation Using MPI Measurements Under Random Losses

Scientific computations are expected to be increasingly distributed across wide-area networks, and Message Passing Interface (MPI) has been shown to scale to support their communications over long distances. Application-level measurements of MPI operations reflect the connection Round-Trip Time (RTT) and loss rate, and machine learning methods have been previously developed to estimate them under deterministic periodic losses. In this paper, we consider more complex, random losses with unform, Poisson and Gaussian distributions. We study five disparate machine leaning methods, with linear and non-linear, and smooth and non-smooth properties, to estimate RTT and loss rate over 10Gbps connections with 0-366ms RTT. The diversity and complexity of these estimators combined with the randomness of losses and TCP’s non-linear response together rule out the selection of a single best among them; instead, we fuse them to retain their design diversity. Overall, the results show that accurate estimates can be generated at low loss rates but become inaccurate at loss rates 10% and higher, thereby illustrating both their strengths and limitations.

Generalization Bounds↗