1,720,977 research outputs found
Parity-violating contributions to nuclear magnetic shielding
No full textThe expressions for both diamagnetic and paramagnetic contributions to the nuclear shielding tensor were reported. Even though both contributions depend on the gauge choice, their sum was gauge independent for sufficiently large basis sets. This formalism was implemented within the framework of ab initio response theory and numerical estimates were reported for two molecules
Parity Violation Energy of Biomolecules - V: Protein Metal Centers
No full textThe parity-violation difference between mirror images of chiral metal centers found in naturally occurring proteins and enzymes is computed at the Dirac-Hartree-Fock level, for both equilibrium and transition state configurations. The systems, selected on the likelihood of yielding high parity violation energies based on atomic mass and coordination geometry, are extracted from: type I Blue Copper Protein active site, Zn and Cd Carbon Anhydrase, Ni Acetyl-Coenzyme-A Synthase, and Mo based CO-Dehydrogenase. Our values provide an approximate upper limit to possible parity-violation effects in biological systems based on static effects
On the Thermodynamic Interpretation of Deep Learning Systems
No full textIn the study of time evolution of the parameters in Deep Learning systems, subject to optimization via SGD (stochastic gradient descent), temperature, entropy and other thermodynamic notions are commonly employed to exploit the Boltzmann formalism. We show that, in simulations on popular databases (CIFAR10, MNIST), such simplified models appear inadequate: different regions in the parameter space exhibit significantly different temperatures and no elementary function expresses the temperature in terms of learning rate and batch size, as commonly assumed. This suggests a more conceptual approach involving contact dynamics and Lie Group Thermodynamics
A novel iterative algorithm to improve segmentations with deep convolutional neural networks trained with synthetic X-ray computed tomography data (i.S.Sy.Da.T.A)
No full textWe propose a novel iterative segmentation algorithm (i.S.Sy.Da.T.A: Iterative Segmentation Synthetic Data Training Algorithm) employing Deep Convolutional Neural Networks and synthetic training data for X-ray tomographic reconstructions of complex microstructures. In our method, we reinforce the synthetic training data with experimental XCT datasets that were automatically segmented in the previous iteration. This strategy produces better segmentations in successive iterations. We test our algorithm with experimental XCT re-constructions of a 6-phase Al-Si Matrix Composite reinforced with ceramic fibers and particles. We perform the analysis in 3D with a special network architecture that demonstrates good generalization with synthetic training data. We show that our iterative algorithm returns better segmentations compared to the standard single training approach. More specifically, phases possessing similar attenuation coefficients can be better segmented: for Al2O3 fibers, SiC particles, and Intermetallics, we see an increase of the Dice score with respect to the classic approach: from 0.49 to 0.54, from 0.66 to 0.72, and from 0.55 to 0.66 respectively. Furthermore, the overall Dice score increases from 0.77 to 0.79. The methods presented in this work are also applicable to other materials and imaging techniques
Topology of magnetic-field-induced current-density field in diatropic monocyclic molecules
No full textConcise information on the general features of the quantum-mechanical current density induced in the
electrons of a molecule by a spatially uniform, time-independent magnetic field is obtained via a stagnation
graph that shows the isolated singularities and the lines at which the current-density vector field vanishes.
Stagnation graphs provide compact description of current-density vector fields and help the interpretation of
molecular magnetic response, e.g., magnetic susceptibility and nuclear magnetic shielding. The stagnation
graph of six cyclic, planar aromatic molecules has been obtained at the Hartree-Fock level via a procedure
based on continuous transformation of the origin of the current density formally annihilating the diamagnetic
contribution. Some common distinctive elements observed for cyclic aromatic rings CnHn, with n
=3,4, . . . ,8, in the presence of a magnetic field normal to the molecular plane, are discussed. The results can
be used for a general discussion of diatropism in aromatic systems
Molecular response to a time-independent non-uniform magnetic-field
No full textThe response of a molecule to a static inhomogeneous magnetic-field is rationalized via multipole magnetic susceptibilities and induced magnetic multipole and anapole moments. The energy of the molecule interacting with the external field is expressed as a Taylor series in the powers of the field and its gradient at the origin of the coordinate system. It involves magnetic multipole tensors of increasing rank, which can be evaluated via quantum mechanical approaches. An electronic energy shift is caused by the feed-back interaction between the induced magnetic dipole moment and the external magnetic field, and between the induced magnetic quadrupole moment and the gradient of the magnetic field. It is shown that, for a static magnetic field with uniform gradient, the magnetic quadrupole moment is origin-dependent, but the total interaction energy and the induced magnetic dipole are invariant to a translation of the coordinate system. The formal advantages of a Geertsen approach to third- and fourth-rank mixed-multipole susceptibilities are discussed. © 2004 Elsevier B.V. All rights reserved
Accelerating NMR Shielding Calculations Through Machine Learning Methods: Application to Magnesium Sodium Silicate Glasses
Full text linkIn this work, we have applied the Kernel Ridge Regression (KRR) method using a Least Square Support Vector Regression (LSSVR) approach for the prediction of the NMR isotropic magnetic shielding (σiso) of active nuclei (17O, 23Na, 25Mg, and 29Si) in a series of (Mg, Na)–silicate glasses. The Machine Learning (ML) algorithm has been trained by mapping the local environment of each atom described by the Smooth Overlap of Atomic Position (SOAP) descriptor with isotropic chemical shielding values computed with DFT using the Gauge-Included-Projector-Augmented-Wave (GIPAW) approach. The influence of different training datasets generated through molecular dynamics simulations at various temperatures and with different inter-atomic potentials has been tested and we demonstrate the importance of a wide exploration of the configurational space to enhance the transferability of the ML-regressor. Finally, the trained ML-regressor has been used to simulate the 29Si MAS NMR spectra of systems containing up to 20000 atoms by averaging hundreds of configurations extracted from classical MD simulations to account for thermal vibrations. This ML approach is a powerful tool for the interpretation of NMR spectra using relatively large systems at a fraction of the computational time required by quantum mechanical calculations which are of high computational cost
A Complete Strategy to Achieve High Precision Automatic Segmentation of Challenging Experimental X-Ray Computed Tomography Data Using Low-Resemblance Synthetic Training Data
Full text linkIt is shown that preconditioning of experimental X-ray computed tomography (XCT) data is critical to achieve high-precision segmentation scores. The challenging experimental XCT datasets and deep convolutional neural networks (DCNNs) are used that are trained with low-resemblance synthetic XCT data. The material used is a 6-phase Al–Si metal matrix composite-reinforced with ceramic fibers and particles. To achieve generalization, in our past studies, specific data augmentation techniques were proposed for the synthetic XCT training data. In addition, two toolsets are devised: (1) special 3D DCNN architecture (3D Triple_UNet), slicing the experimental XCT data from multiple views (MultiView Forwarding), the i.S.Sy.Da.T.A. iterative segmentation algorithm, and (2) nonlocal means (NLM) conditioning (filtering) for the experimental XCT data. This results in good segmentation Dice scores across all phases compared to more standard approaches (i.e., standard UNet architecture, single view slicing, standard single training, and NLM conditioning). Herein, the NLM filter is replaced with the deep conditioning framework BAM SynthCOND introduced in a previous publication, which can be trained with synthetic XCT data. This leads to a significant segmentation precision increase for all phases. The proposed methods are potentially applicable to other materials and imaging techniques
Magnetic response of dithiin molecules: Is there anti-aromaticity in nature?
No full textAb initio current density formalism is used to investigate the response to external magnetic fields of the only known naturally occurring moieties which are formally anti-aromatic, i.e., dithiines. Magnetic susceptibility, nuclear shielding constants, and the topology of induced current densities indicate that although these molecules satisfy Hückel's rule for being anti-aromatic, they are not. In chiral dithiines, the multipolar expansion of the response contains non-vanishing anapole terms associated with 'spinning cuff' current lines. © 2003 Elsevier Science B.V. All rights reserved
Thermodynamics modeling of deep learning systems for a temperature based filter pruning technique
Full text linkWe analyse the dynamics of convolutional filters' parameters of a convolutional neural networks during and after training, via a thermodynamic analogy which allows for a sound definition of temperature. We show that removing high temperature filters has a minor effect on the performance of the model, while removing low temperature filters influences majorly both accuracy and loss decay. This result could be exploited to implement a temperature-based pruning technique for the filters and to determine efficiently the crucial filters for an effective learning
- …
