arXiv subjects

Sanghoon Lee

Publications and source records attributed to Sanghoon Lee.

60 records · Page 4Linked to original sources

arXiv2018

Video Frame Interpolation by Plug-and-Play Deep Locally Linear Embedding

We propose a generative framework which takes on the video frame interpolation problem. Our framework, which we call Deep Locally Linear Embedding (DeepLLE), is powered by a deep convolutional neural network (CNN) while it can be used instantly like conventional models. DeepLLE fits an auto-encoding CNN to a set of several consecutive frames and embeds a linearity constraint on the latent codes so that new frames can be generated by interpolating new latent codes. Different from the current deep learning paradigm which requires training on large datasets, DeepLLE works in a plug-and-play and unsupervised manner, and is able to generate an arbitrary number of frames. Thorough experiments demonstrate that without bells and whistles, our method is highly competitive among current state-of-the-art models.

cs.CV↗

arXiv2018

Unveiling Stimulation Secrets of Electrical Excitation of Neural Tissue Using a Circuit Probability Theory

A new theory, named the Circuit-Probability theory, is proposed to unveil the secret of electrical nerve stimulation, essentially explain the nonlinear and resonant phenomena observed when neural and non-neural tissues are electrically stimulated. For the explanation of frequency dependent response, an inductor is involved in the neural circuit model. Furthermore, predicted response to varied stimulation strength is calculated stochastically. Based on this theory, many empirical models, such as strength-duration relationship and LNP model, can be theoretically explained, derived, and amended. This theory can explain the complex nonlinear interactions in electrical stimulation and fit in vivo experiment data on stimulation-responses of many experiments. As such, the C-P theory should be able to guide novel experiments and more importantly, offer an in-depth physical understanding of the neural tissue. As a promising neural model, we can even further explore the more accurate circuit configuration and probability equation to better describe the electrical stimulation of neural tissues in the future.

q-bio.NC↗

arXiv2012

Critical role of next-nearest-neighbor interlayer interaction in magnetic behavior of magnetic/nonmagnetic multilayers

We report magnetoresistance data in magnetic semiconductor multilayers, which exhibit a clear step-wise behavior as a function of external field. We attribute this highly non-trivial step-wise behavior to next-nearest-neighbor interlayer exchange coupling. Our microscopic calculation suggests that this next-nearest-neighbor coupling can be as large as 24% of the nearest-neighbor coupling. It is argued that such unusually long-range interaction is made possible by the quasi-one-dimensional nature of the system and by the long Fermi wavelength characteristic of magnetic semiconductors.

cond-mat.mtrl-sci↗

arXiv2011

Crossover critical behavior of Ga1-xMnxAs

The critical behavior of Ga1-xMnxAs in a close vicinity of the Curie temperature was experimentally studied by using the thermal diffusivity measurements. Taking into account that the inverse of the thermal diffusivity has the same critical behavior as the specific heat, the critical exponent α for the samples investigated has been determined. With approaching close to the critical temperature, the crossover from the mean-field-like to the Ising-like critical behavior has been observed. From the crossover behavior the values of the Ginzburg number and the exchange interaction length in Ga1-xMnxAs with different concentrations of Mn were determined.

cond-mat.mtrl-sci↗

arXiv2008

Carrier-mediated antiferromagnetic interlayer exchange coupling in diluted magnetic semiconductor multilayers Ga$_{1-x}$Mn$_x$As/GaAs:Be

We use neutron reflectometry to investigate the interlayer exchange coupling between Ga$_{0.97}$Mn$_{0.03}$As ferromagnetic semiconductor layers separated by non-magnetic Be-doped GaAs spacers. Polarized neutron reflectivity measured below the Curie temperature of Ga$_{0.97}$Mn$_{0.03}$As reveals a characteristic splitting at the wave vector corresponding to twice the multilayer period, indicating that the coupling between the ferromagnetic layers are antiferromagnetic (AFM). When the applied field is increased to above the saturation field, this AFM coupling is suppressed. This behavior is not observed when the spacers are undoped, suggesting that the observed AFM coupling is mediated by charge carriers introduced via Be doping. The behavior of magnetization of the multilayers measured by DC magnetometry is consistent with the neutron reflectometry results.

cond-mat.str-el↗

arXiv2008

Distribution of magnetic domain pinning fields in GaMnAs ferromagnetic films

Using the angular dependence of the planar Hall effect in GaMnAs ferromagnetic films, we were able to determine the distribution of magnetic domain pinning fields in this material. Interestingly, there is a major difference between the pinning field distribution in as-grown and in annealed films, the former showing a strikingly narrower distribution than the latter. This conspicuous difference can be attributed to the degree of non-uniformity of magnetic anisotropy in both types of films. This finding provides a better understanding of the magnetic domain landscape in GaMnAs that has been the subject of intense debate.

cond-mat.mtrl-sci↗