The group IV-VI compound tin selenide (SnSe) has recently attracted particular interest due to its unexpectedly low thermal conductivity and high power factor and shows great promise for thermoelectric applications. With an orthorhombic lattice structure, SnSe displays intriguing anisotropic properties due to the low symmetry of the puckered in-plane lattice structure. When thermoelectric materials, such as SnSe, have decreased dimensionality, their thermoelectric conversion efficiency may be improved due to increased power factor and decreased thermal conductivity. Therefore, it is necessary to elucidate the complete optical and electrical anisotropies of SnSe nanostructures in realizing the material's advantages in high-performance devices. Here, we synthesize single-crystal SnSe nanoplates (NPs) using the chemical vapor deposition method. The SnSe NPs' polarized Raman spectra exhibit an angular dependence that reveals the crystal's anomalous anisotropic light-matter interaction. The Raman's anisotropic response has a dependence upon the incident light polarization, photon, and phonon energy, arising from the anisotropic electron-photon and electron-phonon interactions in the SnSe NPs. Finally, angle-resolved charge-transport measurements indicate strong anisotropic conductivity of the SnSe NPs, fully elucidating the anisotropic properties necessary for ultrathin SnSe in electronic, thermoelectric, and optoelectronic devices.
Abstractα-RuCl3 is a major candidate for the realization of the Kitaev quantum spin liquid, but its zigzag antiferromagnetic order at low temperatures indicates deviations from the Kitaev model. We have quantified the spin Hamiltonian of α-RuCl3 by a resonant inelastic x-ray scattering study at the Ru L3 absorption edge. In the paramagnetic state, the quasi-elastic intensity of magnetic excitations has a broad maximum around the zone center without any local maxima at the zigzag magnetic Bragg wavevectors. This finding implies that the zigzag order is fragile and readily destabilized by competing ferromagnetic correlations. The classical ground state of the experimentally determined Hamiltonian is actually ferromagnetic. The zigzag state is stabilized by quantum fluctuations, leaving ferromagnetism – along with the Kitaev spin liquid – as energetically proximate metastable states. The three closely competing states and their collective excitations hold the key to the theoretical understanding of the unusual properties of α-RuCl3 in magnetic fields.
Convolutional neural networks have gained a remarkable success in computer vision. However, most usable network architectures are hand-crafted and usually require expertise and elaborate design. In this paper, we provide a block-wise network generation pipeline called BlockQNN which automatically builds high-performance networks using the Q-Learning paradigm with epsilon-greedy exploration strategy. The optimal network block is constructed by the learning agent which is trained to choose component layers sequentially. We stack the block to construct the whole auto-generated network. To accelerate the generation process, we also propose a distributed asynchronous framework and an early stop strategy. The block-wise generation brings unique advantages: (1) it yields state-of-the-art results in comparison to the hand-crafted networks on image classification, particularly, the best network generated by BlockQNN achieves 2.35% top-1 error rate on CIFAR-10. (2) it offers tremendous reduction of the search space in designing networks, spending only 3 days with 32 GPUs. A faster version can yield a comparable result with only 1 GPU in 20 hours. (3) it has strong generalizability in that the network built on CIFAR also performs well on the larger-scale dataset. The best network achieves very competitive accuracy of 82.0% top-1 and 96.0% top-5 on ImageNet.
Recently, learning equivariant representations has attracted considerable research attention. Dieleman et al. introduce four operations which can be inserted into convolutional neural network to learn deep representations equivariant to rotation. However, feature maps should be copied and rotated four times in each layer in their approach, which causes much running time and memory overhead. In order to address this problem, we propose Deep Rotation Equivariant Network consisting of cycle layers, isotonic layers and decycle layers. Our proposed layers apply rotation transformation on filters rather than feature maps, achieving a speed up of more than 2 times with even less memory overhead. We evaluate DRENs on Rotated MNIST and CIFAR-10 datasets and demonstrate that it can improve the performance of state-of-the-art architectures.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.