-
Bridging Fidelities to Predict Nanoindentation Tip Radii Using Interpretable Deep Learning Models
Authors:
Claus O. W. Trost,
Stanislav Zak,
Sebastian Schaffer,
Christian Saringer,
Lukas Exl,
Megan J. Cordill
Abstract:
As the need for miniaturized structural and functional materials has increased,the need for precise materials characterizaton has also expanded. Nanoindentation is a popular method that can be used to measure material mechanical behavior which enables high-throughput experiments and, in some cases, can also provide images of the indented area through scanning. Both indenting and scanning can cause…
▽ More
As the need for miniaturized structural and functional materials has increased,the need for precise materials characterizaton has also expanded. Nanoindentation is a popular method that can be used to measure material mechanical behavior which enables high-throughput experiments and, in some cases, can also provide images of the indented area through scanning. Both indenting and scanning can cause tip wear that can influence the measurements. Therefore, precise characterization of tip radii is needed to improve data evaluation. A data fusion method is introduced which uses finite element simulations and experimental data to estimate the tip radius in situ in a meaningful way using an interpretable multi-fidelity deep learning approach. By interpreting the machine learning models, it is shown that the approaches are able to accurately capture physical indentation phenomena.
△ Less
Submitted 1 July, 2022;
originally announced July 2022.
-
Material-Dependencies of the THz Emission from Plasmonic Graphene-Based Photoconductive Antenna Structures
Authors:
Christoph Suessmeier,
Sergi Abadal,
Daniel Stock,
Stephan Schaeffer,
Eduard Alarcón,
Seyed Ehsan Hosseininejad,
Anna Katharina Wigger,
Stefan Wagner,
Albert Cabellos-Aparicio,
Max C. Lemme,
Peter Haring Bolívar
Abstract:
Graphene supports surface plasmon polaritons with comparatively slow propagation velocities in the THz region, which becomes increasingly interesting for future communication technologies. This ability can be used to realize compact antennas, which are up to two orders of magnitude smaller than their metallic counterparts. For a proper functionality of these antennas some minimum material requirem…
▽ More
Graphene supports surface plasmon polaritons with comparatively slow propagation velocities in the THz region, which becomes increasingly interesting for future communication technologies. This ability can be used to realize compact antennas, which are up to two orders of magnitude smaller than their metallic counterparts. For a proper functionality of these antennas some minimum material requirements have to be fulfilled, which are presently difficult to achieve, since the fabrication and transfer technologies for graphene are still evolving. In this work we analyze available graphene materials experimentally and extract intrinsic characteristics at THz frequencies, in order to predict the dependency of the THz signal emission threshold as a function of the graphene relaxation time tau_r and the chemical potential mu_c.
△ Less
Submitted 12 April, 2021;
originally announced April 2021.
-
Prediction of magnetization dynamics in a reduced dimensional feature space setting utilizing a low-rank kernel method
Authors:
Lukas Exl,
Norbert J. Mauser,
Sebastian Schaffer,
Thomas Schrefl,
Dieter Suess
Abstract:
We establish a machine learning model for the prediction of the magnetization dynamics as function of the external field described by the Landau-Lifschitz-Gilbert equation, the partial differential equation of motion in micromagnetism. The model allows for fast and accurate determination of the response to an external field which is illustrated by a thin-film standard problem. The data-driven meth…
▽ More
We establish a machine learning model for the prediction of the magnetization dynamics as function of the external field described by the Landau-Lifschitz-Gilbert equation, the partial differential equation of motion in micromagnetism. The model allows for fast and accurate determination of the response to an external field which is illustrated by a thin-film standard problem. The data-driven method internally reduces the dimensionality of the problem by means of nonlinear model reduction for unsupervised learning. This not only makes accurate prediction of the time steps possible, but also decisively reduces complexity in the learning process where magnetization states from simulated micromagnetic dynamics associated with different external fields are used as input data. We use a truncated representation of kernel principal components to describe the states between time predictions. The method is capable of handling large training sample sets owing to a low-rank approximation of the kernel matrix and an associated low-rank extension of kernel principal component analysis and kernel ridge regression. The approach entirely shifts computations into a reduced dimensional setting breaking down the problem dimension from the thousands to the tens.
△ Less
Submitted 19 July, 2021; v1 submitted 13 August, 2020;
originally announced August 2020.
-
Efficient Algorithms for Sampling and Clustering of Large Nonuniform Networks
Authors:
Pekka Orponen,
Satu Elisa Schaeffer
Abstract:
We propose efficient algorithms for two key tasks in the analysis of large nonuniform networks: uniform node sampling and cluster detection. Our sampling technique is based on augmenting a simple, but slowly mixing uniform MCMC sampler with a regular random walk in order to speed up its convergence; however the combined MCMC chain is then only sampled when it is in its "uniform sampling" mode.Ou…
▽ More
We propose efficient algorithms for two key tasks in the analysis of large nonuniform networks: uniform node sampling and cluster detection. Our sampling technique is based on augmenting a simple, but slowly mixing uniform MCMC sampler with a regular random walk in order to speed up its convergence; however the combined MCMC chain is then only sampled when it is in its "uniform sampling" mode.Our clustering algorithm determines the relevant neighbourhood of a given node u in the network by first estimating the Fiedler vector of a Dirichlet matrix with u fixed at zero potential, and then finding the neighbourhood of u that yields a minimal weighted Cheeger ratio, where the edge weights are determined by differences in the estimated node potentials. Both of our algorithms are based on local computations, i.e. operations on the full adjacency matrix of the network are not used. The algorithms are evaluated experimentally using three types of nonuniform networks: Dorogovtsev-Goltsev-Mendes "pseudofractal graphs", scientific collaboration networks, and randomised "caveman graphs".
△ Less
Submitted 2 June, 2004;
originally announced June 2004.
