arXiv subjects

Chen Fang

Publications and source records attributed to Chen Fang.

At least 145 records · Page 8Linked to original sources

arXiv2017

Double-Weyl phonons in transition-metal monosilicides

Topological states of electrons and photons have attracted significant interest recently. Topological mechanical states also being actively explored, have been limited to macroscopic systems of kHz frequency. The discovery of topological phonons of atomic vibrations at THz frequency can provide a new venue for studying heat transfer, phonon scattering and electron-phonon interaction. Here, we employed ab initio calculations to identify a class of noncentrosymmetric materials of $M$Si ($M$=Fe,Co,Mn,Re,Ru) having double Weyl points in both their acoustic and optical phonon spectra. They exhibit novel topological points termed "spin-1 Weyl point" at the Brillouin zone~(BZ) center and "charge-2 Dirac point" at the zone corner. The corresponding gapless surface phonon dispersions are double helicoidal sheets whose isofrequency contours form a single non-contractible loop in the surface BZ. In addition, the global structure of the surface bands can be analytically expressed as double-periodic Weierstrass elliptic functions. Our prediction of topological bulk and surface phonons can be experimentally verified by neutron scattering and electron energy loss spectroscopy, opening brand new directions for topological phononics.

cond-mat.mtrl-sci↗

arXiv2017

Diversified Texture Synthesis with Feed-forward Networks

Recent progresses on deep discriminative and generative modeling have shown promising results on texture synthesis. However, existing feed-forward based methods trade off generality for efficiency, which suffer from many issues, such as shortage of generality (i.e., build one network per texture), lack of diversity (i.e., always produce visually identical output) and suboptimality (i.e., generate less satisfying visual effects). In this work, we focus on solving these issues for improved texture synthesis. We propose a deep generative feed-forward network which enables efficient synthesis of multiple textures within one single network and meaningful interpolation between them. Meanwhile, a suite of important techniques are introduced to achieve better convergence and diversity. With extensive experiments, we demonstrate the effectiveness of the proposed model and techniques for synthesizing a large number of textures and show its applications with the stylization.

cs.CV↗

arXiv2017

Topological crystalline superconductors with linearly and projectively represented $C_{n}$ symmetry

We study superconductors with $n$-fold rotational invariance both in the presence and in the absence of spin-orbit interactions. More specifically, we classify the non-interacting Hamiltonians by defining a series of $Z$-numbers for the Bogoliubov-de Gennes (BdG) symmetry classes of the Altland-Zimbauer classification of random matrices in $1$D, $2$D, and $3$D in the presence of discrete rotational invariance. Our analysis emphasizes the important role played by the angular momentum of the Cooper pairs in the system: for pairings of nonzero angular momentum, the rotation symmetry may be represented projectively, and a projective representation of rotation symmetry may have anomalous properties, including the anti-commutation with the time-reversal symmetry. In 1D and 3D, we show how an $n$-fold axis enhances the topological classification and give additional topological numbers; in 2D, we establish a relation between the Chern number (in class D and CI) and the eigenvalues of rotation symmetry at high-symmetry points. For each nontrivial class in 3D, we write down a minimal effective theory for the surface Majorana states.

cond-mat.supr-con↗

arXiv2016

Topological magnetoplasmon

Classical wave fields are real-valued, ensuring the wave states at opposite frequencies and momenta to be inherently identical. Such a particle-hole symmetry can open up new possibilities for topological phenomena in classical systems. Here we show that the historically studied two-dimensional (2D) magnetoplasmon, which bears gapped bulk states and gapless one-way edge states near zero frequency, is topologically analogous to the 2D topological $p+\Ii p$ superconductor with chiral Majorana edge states and zero modes. We further predict a new type of one-way edge magnetoplasmon at the interface of opposite magnetic domains, and demonstrate the existence of zero-frequency modes bounded at the peripheries of a hollow disk. These findings can be readily verified in experiment, and can greatly enrich the topological phases in bosonic and classical systems.

cond-mat.mes-hall↗

arXiv2016

Scribbler: Controlling Deep Image Synthesis with Sketch and Color

Recently, there have been several promising methods to generate realistic imagery from deep convolutional networks. These methods sidestep the traditional computer graphics rendering pipeline and instead generate imagery at the pixel level by learning from large collections of photos (e.g. faces or bedrooms). However, these methods are of limited utility because it is difficult for a user to control what the network produces. In this paper, we propose a deep adversarial image synthesis architecture that is conditioned on sketched boundaries and sparse color strokes to generate realistic cars, bedrooms, or faces. We demonstrate a sketch based image synthesis system which allows users to 'scribble' over the sketch to indicate preferred color for objects. Our network can then generate convincing images that satisfy both the color and the sketch constraints of user. The network is feed-forward which allows users to see the effect of their edits in real time. We compare to recent work on sketch to image synthesis and show that our approach can generate more realistic, more diverse, and more controllable outputs. The architecture is also effective at user-guided colorization of grayscale images.

cs.CV↗

arXiv2016

Topological nodal line semimetals

We review the recent, mainly theoretical, progress in the study of topological nodal line semimetals in three dimensions. In these semimetals, the conduction and the valence bands cross each other along a one-dimensional curve in the three-dimensional Brillouin zone, and any perturbation that preserves a certain symmetry group (generated by either spatial symmetries or time-reversal symmetry) cannot remove this crossing line and open a full direct gap between the two bands. The nodal line(s) is hence topologically protected by the symmetry group, and can be associated with a topological invariant. In this Review, (i) we enumerate the symmetry groups that may protect a topological nodal line; (ii) we write down the explicit form of the topological invariant for each of these symmetry groups in terms of the wave functions on the Fermi surface, establishing a topological classification; (iii) for certain classes, we review the proposals for the realization of these semimetals in real materials and (iv) we discuss different scenarios that when the protecting symmetry is broken, how a topological nodal line semimetal becomes Weyl semimetals, Dirac semimetals and other topological phases and (v) we discuss the possible physical effects accessible to experimental probes in these materials.

cond-mat.mes-hall↗

arXiv2016

Diagnosis of interaction-driven topological phase via exact diagonalization

We propose a general scheme for diagnosing interaction-driven topological phases in the weak interaction regime using exact diagonalization (ED). The scheme comprises the analysis of eigenvalues of the point-group operators for the many-body eigenstates and the correlation functions for physical observables to extract the symmetries of the order parameters and the topological numbers of the underlying ground states at the thermodynamic limit from a relatively small size system afforded by ED. As a concrete example, we investigate the interaction effects on the half-filled spinless fermions on the checkerboard lattice with a quadratic band crossing point. Numerical results support the existence of a spontaneous quantum anomalous Hall phase purely driven by a nearest-neighbor weak repulsive interaction, separated from a nematic Mott insulator phase at strong repulsive interaction by a first-order phase transition.

cond-mat.str-el↗

arXiv2016

Ultra-small phase estimation via weak measurement technique with postselection

Weak measurement is a novel technique for parameter estimation with higher precision. In this paper we develop a general theory for the parameter estimation based on weak measurement technique with arbitrary postselection. The previous weak value amplification model and the joint weak measurement model are two special cases in our theory. Applying the developed theory, the time-delay estimation is investigated in both theory and experiment. Experimental results shows that when the time-delay is ultra small, the joint weak measurement scheme outperforms the weak value amplification scheme, and is robust against not only the misalignment errors but also the wavelength-dependence of the optical components. These results are consistent with the theoretical predictions that has not been verified by any experiment before.

quant-ph↗

arXiv2016

Vista: A Visually, Socially, and Temporally-aware Model for Artistic Recommendation

Understanding users' interactions with highly subjective content---like artistic images---is challenging due to the complex semantics that guide our preferences. On the one hand one has to overcome `standard' recommender systems challenges, such as dealing with large, sparse, and long-tailed datasets. On the other, several new challenges present themselves, such as the need to model content in terms of its visual appearance, or even social dynamics, such as a preference toward a particular artist that is independent of the art they create. In this paper we build large-scale recommender systems to model the dynamics of a vibrant digital art community, Behance, consisting of tens of millions of interactions (clicks and `appreciates') of users toward digital art. Methodologically, our main contributions are to model (a) rich content, especially in terms of its visual appearance; (b) temporal dynamics, in terms of how users prefer `visually consistent' content within and across sessions; and (c) social dynamics, in terms of how users exhibit preferences both towards certain art styles, as well as the artists themselves.

cs.IR↗

arXiv2016

Co-existence of Weyl Fermion and Massless Triply Degenerate Nodal Points

By using first-principles calculations, we propose that WC-type ZrTe is a new type of topological semimetal (TSM). It has six pairs of chiral Weyl nodes in its first Brillouin zone, but it is distinguished from other existing TSMs by having additional two paris of massless fermions with triply degenerate nodal points as proposed in the isostructural compounds TaN and NbN. The mirror symmetry, three-fold rotational symmetry and time-reversal symmetry require all of the Weyl nodes to have the same velocity vectors and locate at the same energy level. The Fermi arcs on different surfaces are shown, which may be measured by future experiments. It demonstrates that the "material universe" can support more intriguing particles simultaneously.

cond-mat.mes-hall↗

arXiv2016

Van der Waals Stacking Induced Topological Phase Transition in Layered Ternary Transition Metal Chalcogenides

Novel materials with nontrivial electronic and photonic band topology are crucial for realizing novel devices with low power consumption and heat dissipation, and quantum computing free of decoherence. Here using first-principles approach, we predict a class of ternary transition metal chalcogenides (TTMC) MM'Te$_4$ exhibits dual topological characteristics: quantum spin Hall (QSH) insulators in their 2D monolayers and topological Weyl semimetals in their 3D noncentrosymmetric crystals upon van der Waals (vdW) stacking. Remarkably, we find that one can create and annihilate Weyl fermions, and realize the transition between Type-I and Type-II Weyl fermions by tuning vdW interlayer spacing. Our calculations show that they possess excellent thermodynamic stability and weak interlayer binding, implying their great potentials for experimental synthesis, direct exfoliation and vdW heterostacking. Moreover, their ternary nature will offer more tunability for electronic structure by controlling different stoichiometry and valence charges. Our findings provide an ideal materials platform for realizing QSH effect and exploring topological phase transition, and will open up a variety of new opportunities for two-dimensional materials and topological materials research.

cond-mat.mtrl-sci↗

arXiv2016

Topological Semimetals with Triply Degenerate Nodal Points in θ-phase Tantalum Nitride

Using first-principles calculation and symmetry analysis, we propose that θ-TaN is a topological semimetal having a new type of point nodes, i.e., triply degenerate nodal points. Each node is a band crossing between degenerate and non-degenerate bands along the high-symmetry line in the Brillouin zone, and is protected by crystalline symmetries. Such new type of nodes will always generate singular touching points between different Fermi surfaces and 3D spin texture around them. Breaking the crystalline symmetry by external magnetic field or strain leads to various of topological phases. By studying the Landau levels under a small field along $c$-axis, we demonstrate that the system has a new quantum anomaly that we call "helical anomaly".

cond-mat.mtrl-sci↗

arXiv2016

Tunable Weyl fermions and Fermi arcs in magnetized topological crystalline insulators

Based on $k\cdot p$ analysis and realistic tight-binding calculations, we find that time-reversal-breaking Weyl semimetals can be realized in magnetically-doped (Mn, Eu, Cr etc.) Sn$_{1-x}$Pb$_x$(Te,Se) class of topological crystalline insulators. All the Weyl points are well separated in momentum space and possess nearly the same energy due to high crystalline symmetry. Moreover, both the Weyl points and Fermi arcs are highly tunable by varying Pb/Sn composition, pressure, magnetization, temperature, surface potential etc., opening up the possibility of manipulating Weyl points and rewiring the Fermi arcs.

cond-mat.mes-hall↗

arXiv2016

Modular Anomalies in (2+1) and (3+1)-D Edge Theories

The classification of topological phases of matter in the presence of interactions is an area of intense interest. One possible means of classification is via studying the partition function under modular transforms, as the presence of an anomalous phase arising in the edge theory of a D-dimensional system under modular transformation, or modular anomaly, signals the presence of a (D+1)-D non-trivial bulk. In this work, we discuss the modular transformations of conformal field theories along a (2+1)-D and a (3+1)-D edge. Using both analytical and numerical methods, we show that chiral complex free fermions in (2+1)-D and (3+1)-D are modular invariant. However, we show in (3+1)-D that when the edge theory is coupled to a background U(1) gauge field this results in the presence of a modular anomaly that is the manifestation of a quantum Hall effect in a (4+1)-D bulk. Using the modular anomaly, we find that the edge theory of (4+1)-D insulator with spacetime inversion symmetry(P*T) and fermion number parity symmetry for each spin becomes modular invariant when 8 copies of the edges exist.

cond-mat.str-el↗

arXiv2016

Image Captioning with Semantic Attention

Automatically generating a natural language description of an image has attracted interests recently both because of its importance in practical applications and because it connects two major artificial intelligence fields: computer vision and natural language processing. Existing approaches are either top-down, which start from a gist of an image and convert it into words, or bottom-up, which come up with words describing various aspects of an image and then combine them. In this paper, we propose a new algorithm that combines both approaches through a model of semantic attention. Our algorithm learns to selectively attend to semantic concept proposals and fuse them into hidden states and outputs of recurrent neural networks. The selection and fusion form a feedback connecting the top-down and bottom-up computation. We evaluate our algorithm on two public benchmarks: Microsoft COCO and Flickr30K. Experimental results show that our algorithm significantly outperforms the state-of-the-art approaches consistently across different evaluation metrics.

cs.CV↗

arXiv2015

Multi-Instance Visual-Semantic Embedding

Visual-semantic embedding models have been recently proposed and shown to be effective for image classification and zero-shot learning, by mapping images into a continuous semantic label space. Although several approaches have been proposed for single-label embedding tasks, handling images with multiple labels (which is a more general setting) still remains an open problem, mainly due to the complex underlying corresponding relationship between image and its labels. In this work, we present Multi-Instance visual-semantic Embedding model (MIE) for embedding images associated with either single or multiple labels. Our model discovers and maps semantically-meaningful image subregions to their corresponding labels. And we demonstrate the superiority of our method over the state-of-the-art on two tasks, including multi-label image annotation and zero-shot learning.

cs.CV↗

arXiv2015

Topological semimetals with Riemann surface states

Riemann surfaces are geometric constructions in complex analysis that may represent multi-valued holomorphic functions using multiple sheets of the complex plane. We show that the energy dispersion of surface states in topological semimetals can be represented by Riemann surfaces generated by holomorphic functions in the two-dimensional momentum space, whose constant height contours correspond to Fermi arcs. This correspondence is demonstrated in the recently discovered Weyl semimetals and leads us to predict new types of topological semimetals, whose surface states are represented by double- and quad-helicoid Riemann surfaces. The intersection of multiple helicoids, or the branch cut of the generating function, appears on high-symmetry lines in the surface Brillouin zone, where surface states are guaranteed to be doubly degenerate by a glide reflection symmetry. We predict the heterostructure superlattice [(SrIrO$_3$)$_2$(CaIrO$_3$)$_2$] to be a topological semimetal with double-helicoid Riemann surface states.

cond-mat.mes-hall↗

arXiv2015

Symmetry fractionalization of visons in $\mathbb Z_2$ spin liquids

In this work we study symmetry fractionalization of vison excitations in topological $\mathbb{Z}_2$ spin liquids. We show that in the presence of the full $\mathrm{SO}(3)$ spin rotational symmetry and if there is an odd number of spin-$\frac12$ per unit cell, the symmetry fractionalization of visons is completely fixed. On the other hand, visons can have different classes of symmetry fractionalization if the spin rotational symmetry is reduced. As a concrete example, we show that visons in the Balents-Fisher-Girvin $\mathbb{Z}_2$ spin liquid have crystal symmetry fractionalization classes which are not allowed in $\mathrm{SO}(3)$ symmetric spin liquids, due to the reduced spin rotational symmetry.

cond-mat.str-el↗