arXiv subjects

Ji Feng

Publications and source records attributed to Ji Feng.

At least 37 records · Page 2Linked to original sources

arXiv2022

Emergence of bond-dependent highly anisotropic magnetic interactions in Sr$_4$RhO$_6$: a theoretical study

The quantum spin liquid states as a natural ground state of the Kitaev model has led to a quest for new materials candidates hosting Kitaev physics. Yet, there are very few material candidates in this category. Using a combination of $ab$ $initio$ and model Hamiltonian methods, we propose that Ruddlesden-Popper compound Sr$_4$RhO$_6$ belongs to this category. With a tight-binding model and exact diagonalization approach, we show that despite substantial trigonal-like distortion, the electronic and magnetic properties of Sr$_4$RhO$_6$ can be well described in terms of pseudo-spin = 1/2 states. Magnetic interactions among pseudo-spins, estimated using the second-order perturbation method are highly bond-dependent anisotropic in nature with two particularly noticeable features, antiferromagnetic Kitaev and Dzyaloshinskii-Moriya interactions. The gaped spin-wave spectra of Sr$_4$RhO$_6$ obtained with linear spin-wave theory is consistent with the underlying magnetic frustration. Additional analysis of the role of individual or a particular combination of magnetic interactions reveals that the spin-wave spectra of Sr$_4$RhO$_6$ is a combined effect of the highly anisotropic interactions and a relatively simpler minimal model may not be plausible in the current case. The crucial insights about coupling between the local structural features and magnetic properties of Sr$_4$RhO$_6$ obtained in this study may be helpful for future studies belonging to this class.

cond-mat.str-el↗

arXiv2022

Spin interaction and magnetism in cobaltate Kitaev candidate materials: an $ab$ $initio$ and model Hamiltonian approach

In the quest for materials hosting Kitaev spin liquids, much of the efforts have been focused on the fourth- and fifth-row transition metal compounds, which are spin-orbit coupling assisted Mott insulators. Here, we study the structural and magnetic properties of 3$d$ transition metal oxides, Na$_2$Co$_2$TeO$_6$ and Na$_3$Co$_2$SbO$_6$. The partial occupancy of sodium in former compound is addressed using a cluster expansion, and a honeycomb lattice of sodiums is found to be energetically favored. Starting from the \textit{ab initio} band structures, a many-body second order perturbation theory leads to a pseudospin-$\frac{1}{2}$ Hamiltonian with estimated magnetic interactions. We show that the experimentally observed zigzag magnetic state is stabilized only when the first neighbor Kitaev coupling dominates over the Heisenberg term, both of which are highly suppressed due to presence of $e_g$ orbitals. A third neighbor Heisenberg interaction is found dominant in both these compounds. We also present a phase diagram for Na$_2$Co$_2$TeO$_6$ by varying the electron-electron and spin-orbit interactions. The computed spin excitation spectra are found to capture essential features of recent experimental magnon spectrum, lending support to our results.

cond-mat.str-el↗

arXiv2022

Neural network representation of electronic structure from $ab$ $initio$ molecular dynamics

Despite their rich information content, electronic structure data amassed at high volumes in $ab$ $initio$ molecular dynamics simulations are generally under-utilized. We introduce a transferable high-fidelity neural network representation of such data in the form of tight-binding Hamiltonians for crystalline materials. This predictive representation of $ab$ $initio$ electronic structure, combined with machine-learning boosted molecular dynamics, enables efficient and accurate electronic evolution and sampling. When applied to a one-dimension charge-density wave material, carbyne, we are able to compute the spectral function and optical conductivity in the canonical ensemble. The spectral functions evaluated during soliton-antisoliton pair annihilation process reveal significant renormalization of low-energy edge modes due to retarded electron-lattice coupling beyond the Born-Oppenheimer limit. The availability of an efficient and reusable surrogate model for the electronic structure dynamical system will enable calculating many interesting physical properties, paving way to previously inaccessible or challenging avenues in materials modeling.

cond-mat.mtrl-sci↗

arXiv2022

Topological surfaces of domain wall-decorated antiferromagnetic topological insulator MnBi$_{2n}$Te$_{3n+1}$

Antiferromagnetic topological insulators harbor topological in-gap surface states protected by an anti-unitary $S$ symmetry, which is broken by the inevitable presence of domain walls. Whether an antiferromagnetic topological insulator with domain walls is gapless and metallic on its topological surfaces remains to be elucidated. We show that a single non-statistical index characterizing the magnetic order of domain wall-decorated antiferromagnetic topological insulator, referred to as the Ising moment, determines the topological surface gap, which can be zero even when the $S$ symmetry is manifestly broken. In the thermodynamic limit, the topological surface states tend to be gapless when magnetic fluctuation is bounded. In this case, the Lyapunov exponent of the surface transfer matrix reveals a surface delocalization transition near the zero energy due to a crossover from orthogonal to chiral orthogonal symmetry class. Spectroscopic and transport measurements on the surface states will reveal the critical behavior of the transition, which in return bears on the nature of antiferromagnetic domains walls.

cond-mat.mes-hall↗

arXiv2022

Charge doping to flat AgF2 monolayers in a chemical capacitor setup

Flat monolayers of silver II fluoride, which could be obtained by epitaxial deposition on an appropriate substrate, have been recently predicted to exhibit very strong antiferromagnetic superexchange and to have large potential for ambient pressure superconductivity if doped to an optimal level. It was shown that AgF2 could become a magnetic glue based superconductor with a critical superconducting temperature approaching 200 K at optimum doping. In the current work we calculate the optimum doping to correspond to 14% of holes per formula unit, i.e. quite similar to that for oxocuprates II. Furthermore, using DFT calculations we show that flat AgF2 single layers can indeed be doped to a controlled extent using a recently proposed chemical capacitor setup. Hole doping associated with formation of Ag III proves to be difficult to achieve in the setup explored in this work as it falls at verge of charge stability of fluoride anions and does not affect the d x2 minus y2 manifold . However, in the case of electron doping, manipulation of different factors, such as number of dopant layers and the thickness of the separator, permits fine tuning of the doping level - and concomitantly TC - all the way from underdoped to overdoped regime (in a similar manner as chemical doping for the Nd2CuO4 analogue).

cond-mat.supr-con↗

arXiv2021

Worldline algorithm by oracle-guided variational autoregressive network

The variational autoregressive network is extended to the Euclidean path integral representation of quantum partition function. An essential challenge is adapting the sequential process of sample generation by an autoregressive network to a nonlocal constraint due to the periodic boundary condition in path integral. An add-on oracle is devised for this purpose, which accurately identifies and stalls unviable configurations as soon as they occur. The oracle enables rejection-free sampling conforming to the periodic boundary condition. As a demonstration, the oracle-guided autoregressive network is applied to obtain variational solutions of quantum spin chains at finite temperatures with relatively large system sizes and numbers of time slicing, and to efficiently compute thermodynamic quantities.

cond-mat.stat-mech↗

arXiv2021

A silver(II) route to unconventional superconductivity

The highly unusual divalent silver in silver difluoride (AgF$_2$) features a nearly square lattice of Ag$^{+2}$ bridged by fluorides. As a structural and electronic analogue of cuprates, its superconducting properties are yet to be examined. Our first principles electronic structure calculations reveal a striking resemblance between AgF$_2$ and the cuprates. Computed spin susceptibility shows a magnetic instability consistent with the experimentally observed antiferromagnetic transition. A linearized Eliashberg theory in fluctuation-exchange approximation shows an unconventional singlet $d$-wave superconducting pairing for bulk AgF$_2$ at an optimal electron doping. The pairing is found to strengthen with a decreasing interlayer coupling, highlighting the importance of quasi-2D nature of the crystal structure. These findings place AgF$_2$ in the category of unconventional high-$T_\text{C}$ superconductors, and its chemical uniqueness may help shed new lights on the high-$T_\text{C}$ phenomena.

cond-mat.supr-con↗

arXiv2020

Dynamical correlation enhanced orbital magnetization in VI$_{3}$

The effect of electronic correlations on the orbital magnetization in real materials has not been explored beyond a static mean-field level. Based on the dynamical mean-field theory, the effect of electronic correlations on the orbital magnetization in layered ferromagnet VI$_3$ has been studied. A comparison drawn with the results obtained from density functional theory calculations robustly establishes the crucial role of dynamical correlations in this case. In contrast to the density functional theory that leads to negligible orbital magnetization in VI$_3$, in dynamical mean-field approach the orbital magnetization is greatly enhanced. Further analysis show that this enhancement is mainly due to the enhanced local circulations of electrons, which can be attributed to a better description of the localization behavior of correlated electrons in VI$_3$. The conclusion drawn in our study could be applicable to a wide range of layered materials in this class.

cond-mat.str-el↗

arXiv2020

Deep Forest

Current deep learning models are mostly build upon neural networks, i.e., multiple layers of parameterized differentiable nonlinear modules that can be trained by backpropagation. In this paper, we explore the possibility of building deep models based on non-differentiable modules. We conjecture that the mystery behind the success of deep neural networks owes much to three characteristics, i.e., layer-by-layer processing, in-model feature transformation and sufficient model complexity. We propose the gcForest approach, which generates \textit{deep forest} holding these characteristics. This is a decision tree ensemble approach, with much less hyper-parameters than deep neural networks, and its model complexity can be automatically determined in a data-dependent way. Experiments show that its performance is quite robust to hyper-parameter settings, such that in most cases, even across different data from different domains, it is able to get excellent performance by using the same default setting. This study opens the door of deep learning based on non-differentiable modules, and exhibits the possibility of constructing deep models without using backpropagation.

cs.LG↗

arXiv2020

Eightfold Fermionic Excitation in a Charge Density Wave Compound

Unconventional quasiparticle excitations in condensed matter systems have become one of the most important research frontiers. Beyond two- and fourfold degenerate Weyl and Dirac fermions, three-, six- and eightfold symmetry protected degeneracies have been predicted however remain challenging to realize in solid state materials. Here, charge density wave compound TaTe4 is proposed to hold eightfold fermionic excitation and Dirac point in energy bands. High quality TaTe4 single crystals are prepared, where the charge density wave is revealed by directly imaging the atomic structure and a pseudogap of about 45 meV on the surface. Shubnikov de-Haas oscillations of TaTe4 are consistent with band structure calculation. Scanning tunneling microscopy reveals atomic step edge states on the surface of TaTe4. This work uncovers that charge density wave is able to induce new topological phases and sheds new light on the novel excitations in condensed matter materials.

cond-mat.mtrl-sci↗

arXiv2020

Soft Gradient Boosting Machine

Gradient Boosting Machine has proven to be one successful function approximator and has been widely used in a variety of areas. However, since the training procedure of each base learner has to take the sequential order, it is infeasible to parallelize the training process among base learners for speed-up. In addition, under online or incremental learning settings, GBMs achieved sub-optimal performance due to the fact that the previously trained base learners can not adapt with the environment once trained. In this work, we propose the soft Gradient Boosting Machine (sGBM) by wiring multiple differentiable base learners together, by injecting both local and global objectives inspired from gradient boosting, all base learners can then be jointly optimized with linear speed-up. When using differentiable soft decision trees as base learner, such device can be regarded as an alternative version of the (hard) gradient boosting decision trees with extra benefits. Experimental results showed that, sGBM enjoys much higher time efficiency with better accuracy, given the same base learner in both on-line and off-line settings.

cs.LG↗

arXiv2020

A non-iterative method for the vertex corrections of the Kubo formula for electric conductivity

In computing electric conductivity based on the Kubo formula, the vertex corrections describe such effects as anisotropic scattering and quantum interference and are important to quantum transport properties. These vertex corrections are obtained by solving Bethe-Salpeter equations, which can become numerically intractable when a large number of k-points and multiple bands are involved. We introduce a non-iterative approach to the vertex correction based on rank factorization of the impurity vertices, which significantly alleviate the computational burden. We demonstrate that this method can be implemented along with effective Hamiltonians extracted from electronic structure calculations on perfect crystals, thereby enabling quantitative analysis of quantum effects in electron conduction for real materials.

cond-mat.mtrl-sci↗

arXiv2020

Distributed Deep Forest and its Application to Automatic Detection of Cash-out Fraud

Internet companies are facing the need for handling large-scale machine learning applications on a daily basis and distributed implementation of machine learning algorithms which can handle extra-large scale tasks with great performance is widely needed. Deep forest is a recently proposed deep learning framework which uses tree ensembles as its building blocks and it has achieved highly competitive results on various domains of tasks. However, it has not been tested on extremely large scale tasks. In this work, based on our parameter server system, we developed the distributed version of deep forest. To meet the need for real-world tasks, many improvements are introduced to the original deep forest model, including MART (Multiple Additive Regression Tree) as base learners for efficiency and effectiveness consideration, the cost-based method for handling prevalent class-imbalanced data, MART based feature selection for high dimension data and different evaluation metrics for automatically determining of the cascade level. We tested the deep forest model on an extra-large scale task, i.e., automatic detection of cash-out fraud, with more than 100 millions of training samples. Experimental results showed that the deep forest model has the best performance according to the evaluation metrics from different perspectives even with very little effort for parameter tuning. This model can block fraud transactions in a large amount of money each day. Even compared with the best-deployed model, the deep forest model can additionally bring into a significant decrease in economic loss each day.

cs.LG↗

arXiv2020

Giant pressure-enhancement of multiferroicity in CuBr2

Type-II multiferroic materials, in which ferroelectric polarization is induced by inversion non-symmetric magnetic order, promise new and highly efficient multifunctional applications based on the mutual control of magnetic and electric properties. Although this phenomenon has to date been limited to low temperatures, here we report a giant pressure-dependence of the multiferroic critical temperature in CuBr$_2$. At 4.5 GPa, $T_\mathrm{C}$ is enhanced from 73.5 to 162 K, to our knowledge the highest value yet reported for a non-oxide type-II multiferroic. This growth shows no sign of saturating and the dielectric loss remains small under these high pressures. We establish the structure under pressure and demonstrate a 60\% increase in the two-magnon Raman energy scale up to 3.6 GPa. First-principles structural and magnetic energy calculations provide a quantitative explanation in terms of dramatically pressure-enhanced interactions between CuBr$_2$ chains. These large, pressure-tuned magnetic interactions motivate structural control in cuprous halides as a route to applied high-temperature multiferroicity.

cond-mat.mtrl-sci↗

arXiv2019

Emergent superconductivity in two-dimensional NiTe$_2$ crystals

Despite growing interest in them, highly crystalline two-dimensional superconductors derived from exfoliated layered materials are few. Employing the anisotropic Migdal-Eliashberg formalism based on {\it ab initio} calculations, we find monolayer NiTe$_{2}$ to be an intrinsic superconductor with a $T_{\text c}\sim$5.7~K, although the bulk crystal is not known to superconduct. Remarkably, bilayer NiTe$_{2}$ intercalated with lithium is found to display two-gap superconductivity with a critical temperature $T_{\text{c}}\sim 11.3$~K and superconducting gap of $\sim$3.1~meV, arising from a synergy of electronic and phononic effects. The comparatively high $T_\text{c}$, substrate independence and proximity tunability will make these superconductors ideal platforms for exploring intriguing correlation effects and quantum criticality associated two-dimensional superconductivity.

cond-mat.supr-con↗

arXiv2019

Learning to Confuse: Generating Training Time Adversarial Data with Auto-Encoder

In this work, we consider one challenging training time attack by modifying training data with bounded perturbation, hoping to manipulate the behavior (both targeted or non-targeted) of any corresponding trained classifier during test time when facing clean samples. To achieve this, we proposed to use an auto-encoder-like network to generate the pertubation on the training data paired with one differentiable system acting as the imaginary victim classifier. The perturbation generator will learn to update its weights by watching the training procedure of the imaginary classifier in order to produce the most harmful and imperceivable noise which in turn will lead the lowest generalization power for the victim classifier. This can be formulated into a non-linear equality constrained optimization problem. Unlike GANs, solving such problem is computationally challenging, we then proposed a simple yet effective procedure to decouple the alternating updates for the two networks for stability. The method proposed in this paper can be easily extended to the label specific setting where the attacker can manipulate the predictions of the victim classifiers according to some predefined rules rather than only making wrong predictions. Experiments on various datasets including CIFAR-10 and a reduced version of ImageNet confirmed the effectiveness of the proposed method and empirical results showed that, such bounded perturbation have good transferability regardless of which classifier the victim is actually using on image data.

cs.LG↗

arXiv2019

Distinct fingerprints of charge density waves and electronic standing waves in ZrTe$_3$

Experimental signatures of charge density waves (CDW) in high-temperature superconductors have evoked much recent interest, yet an alternative interpretation has been theoretically raised based on electronic standing waves resulting from quasiparticles scattering off impurities or defects, also known as Friedel oscillations (FO). Indeed the two phenomena are similar and related, posing a challenge to their experimental differentiation. Here we report a resonant X-ray diffraction study of ZrTe$_3$, a model CDW material. Near the CDW transition, we observe two independent diffraction signatures that arise concomitantly, only to become clearly separated in momentum while developing very different correlation lengths in the well-ordered state. Anomalously slow dynamics of mesoscopic ordered nanoregions are further found near the transition temperature, in spite of the expected strong thermal fluctuations. These observations reveal that a spatially-modulated CDW phase emerges out of a uniform electronic fluid via a process that is promoted by self-amplifying FO, and identify a viable experimental route to distinguish CDW and FO.

cond-mat.str-el↗

arXiv2019

High-pressure lithium as an elemental topological semimetal

Topological semimetals generally contain heavy elements. Using density-functional theoretic calculations, we predict that three dense lithium polymorphs in the pressure range 200--360 GPa display nontrivial semimetallic electronic structure. Specifically, these high-pressure phases exhibit Fermi pockets which are degenerate over a loop in $\boldsymbol{k}$-space, around which an encircling $\bm k$-space path is threaded by $\pm π$ Berry phase. Accordingly, these dense lithium phases are topological nodal loop semimetals involving a single light element.

cond-mat.mtrl-sci↗