-
Charge order induced Dirac pockets in the nonsymmorphic crystal TaTe$_4$
Authors:
Yichen Zhang,
Ruixiang Zhou,
Hanlin Wu,
Ji Seop Oh,
Sheng Li,
Jianwei Huang,
Jonathan D. Denlinger,
Makoto Hashimoto,
Donghui Lu,
Sung-Kwan Mo,
Kevin F. Kelly,
Gregory T. McCandless,
Julia Y. Chan,
Robert J. Birgeneau,
Bing Lv,
Gang Li,
Ming Yi
Abstract:
The interplay between charge order (CO) and nontrivial band topology has spurred tremendous interest in understanding topological excitations beyond the single-particle description. In a quasi-one-dimensional nonsymmorphic crystal TaTe$_4$, the (2a$\times$2b$\times$3c) charge ordered ground state drives the system into a space group where the symmetry indicator features the emergence of Dirac ferm…
▽ More
The interplay between charge order (CO) and nontrivial band topology has spurred tremendous interest in understanding topological excitations beyond the single-particle description. In a quasi-one-dimensional nonsymmorphic crystal TaTe$_4$, the (2a$\times$2b$\times$3c) charge ordered ground state drives the system into a space group where the symmetry indicator features the emergence of Dirac fermions and unconventional double Dirac fermions. Using angle-resolved photoemission spectroscopy and first-principles calculations, we provide evidence of the CO induced Dirac fermion-related bands near the Fermi level. Furthermore, the band folding at the Fermi level is compatible with the new periodicity dictated by the CO, indicating that the electrons near the Fermi level follow the crystalline symmetries needed to host double Dirac fermions in this system.
△ Less
Submitted 25 March, 2024; v1 submitted 1 April, 2023;
originally announced April 2023.
-
Compressed Domain Image Classification Using a Dynamic-Rate Neural Network
Authors:
Yibo Xu,
Weidi Liu,
Kevin F. Kelly
Abstract:
Compressed domain image classification performs classification directly on compressive measurements acquired from the single-pixel camera, bypassing the image reconstruction step. It is of great importance for extending high-speed object detection and classification beyond the visible spectrum in a cost-effective manner especially for resource-limited platforms. Previous neural network methods req…
▽ More
Compressed domain image classification performs classification directly on compressive measurements acquired from the single-pixel camera, bypassing the image reconstruction step. It is of great importance for extending high-speed object detection and classification beyond the visible spectrum in a cost-effective manner especially for resource-limited platforms. Previous neural network methods require training a dedicated neural network for each different measurement rate (MR), which is costly in computation and storage. In this work, we develop an efficient training scheme that provides a neural network with dynamic-rate property, where a single neural network is capable of classifying over any MR within the range of interest with a given sensing matrix. This training scheme uses only a few selected MRs for training and the trained neural network is valid over the full range of MRs of interest. We demonstrate the performance of the dynamic-rate neural network on datasets of MNIST, CIFAR-10, Fashion-MNIST, COIL-100, and show that it generates approximately equal performance at each MR as that of a single-rate neural network valid only for one MR. Robustness to noise of the dynamic-rate model is also demonstrated. The dynamic-rate training scheme can be regarded as a general approach compatible with different types of sensing matrices, various neural network architectures, and is a valuable step towards wider adoption of compressive inference techniques and other compressive sensing related tasks via neural networks.
△ Less
Submitted 14 December, 2020; v1 submitted 28 January, 2019;
originally announced January 2019.
-
A Hardware Realization of Superresolution Combining Random Coding and Blurring
Authors:
Kevin Beale,
Jianbo Chen,
Kevin F. Kelly,
Justin Romberg
Abstract:
Resolution enhancements are often desired in imaging applications where high-resolution sensor arrays are difficult to obtain. Many computational imaging methods have been proposed to encode high-resolution scene information on low-resolution sensors by cleverly modulating light from the scene before it hits the sensor. These methods often require movement of some portion of the imaging apparatus…
▽ More
Resolution enhancements are often desired in imaging applications where high-resolution sensor arrays are difficult to obtain. Many computational imaging methods have been proposed to encode high-resolution scene information on low-resolution sensors by cleverly modulating light from the scene before it hits the sensor. These methods often require movement of some portion of the imaging apparatus or only acquire images up to the resolution of a modulating element. Here a technique is presented for resolving beyond the resolutions of both a pointwise-modulating mask element and a sensor array through the introduction of a controlled blur into the optical pathway. The analysis contains an intuitive and exact expression for the overall superresolvability of the system, and arguments are presented to explain how the combination of random coding and blurring makes the superresolution problem well-posed. Experimental results demonstrate that a resolution enhancement of approximately $4\times$ is possible in practice using standard optical components, without mechanical motion of the imaging apparatus, and without any a priori assumptions on scene structure.
△ Less
Submitted 20 October, 2018;
originally announced October 2018.
-
Topological metal behavior in GeBi2Te4 single crystals
Authors:
A. Marcinkova,
J. K. Wang,
C. Slavonic,
Andriy H. Nevidomskyy,
K. F. Kelly,
Y. Filinchuk,
E. Morosan
Abstract:
The metallic character of the GeBi2Te4 single crystals is probed using a combination of structural and physical properties measurements, together with density functional theory (DFT) calculations. The structural study shows distorted Ge coordination polyhedra, mainly of the Ge octahedra. This has a major impact on the band structure, resulting in bulk metallic behavior of GeBi2Te4, as indicated by…
▽ More
The metallic character of the GeBi2Te4 single crystals is probed using a combination of structural and physical properties measurements, together with density functional theory (DFT) calculations. The structural study shows distorted Ge coordination polyhedra, mainly of the Ge octahedra. This has a major impact on the band structure, resulting in bulk metallic behavior of GeBi2Te4, as indicated by DFT calculations. Such calculations place GeBi2Te4 in a class of a few known non-trivial topological metals, and explains why an observed Dirac point lies below the Fermi energy at about -0.12eV. A topological picture of GeBi2Te4 is confirmed by the observation of surface state modulations by scanning tunneling microscopy (STM).
△ Less
Submitted 10 January, 2014;
originally announced January 2014.
-
The STONE Transform: Multi-Resolution Image Enhancement and Real-Time Compressive Video
Authors:
Tom Goldstein,
Lina Xu,
Kevin F. Kelly,
Richard Baraniuk
Abstract:
Compressed sensing enables the reconstruction of high-resolution signals from under-sampled data. While compressive methods simplify data acquisition, they require the solution of difficult recovery problems to make use of the resulting measurements. This article presents a new sensing framework that combines the advantages of both conventional and compressive sensing. Using the proposed \stone tr…
▽ More
Compressed sensing enables the reconstruction of high-resolution signals from under-sampled data. While compressive methods simplify data acquisition, they require the solution of difficult recovery problems to make use of the resulting measurements. This article presents a new sensing framework that combines the advantages of both conventional and compressive sensing. Using the proposed \stone transform, measurements can be reconstructed instantly at Nyquist rates at any power-of-two resolution. The same data can then be "enhanced" to higher resolutions using compressive methods that leverage sparsity to "beat" the Nyquist limit. The availability of a fast direct reconstruction enables compressive measurements to be processed on small embedded devices. We demonstrate this by constructing a real-time compressive video camera.
△ Less
Submitted 15 November, 2013; v1 submitted 14 November, 2013;
originally announced November 2013.
-
Interfacial charge transfer in nanoscale polymer transistors
Authors:
J. H. Worne,
R. Giridharagopal,
K. F. Kelly,
D. Natelson
Abstract:
Interfacial charge transfer plays an essential role in establishing the relative alignment of the metal Fermi level and the energy bands of organic semiconductors. While the details remain elusive in many systems, this charge transfer has been inferred in a number of photoemission experiments. We present electronic transport measurements in very short channel ($L < 100$ nm) transistors made from…
▽ More
Interfacial charge transfer plays an essential role in establishing the relative alignment of the metal Fermi level and the energy bands of organic semiconductors. While the details remain elusive in many systems, this charge transfer has been inferred in a number of photoemission experiments. We present electronic transport measurements in very short channel ($L < 100$ nm) transistors made from poly(3-hexylthiophene) (P3HT). As channel length is reduced, the evolution of the contact resistance and the zero-gate-voltage conductance are consistent with such charge transfer. Short channel conduction in devices with Pt contacts is greatly enhanced compared to analogous devices with Au contacts, consistent with charge transfer expectations. Alternating current scanning tunneling microscopy (ACSTM) provides further evidence that holes are transferred from Pt into P3HT, while much less charge transfer takes place at the Au/P3HT interface.
△ Less
Submitted 22 September, 2008;
originally announced September 2008.
