arXiv subjects

Ci Li

Publications and source records attributed to Ci Li.

17 recordsLinked to original sources

arXiv2025

Linear exciton Hall and Nernst effects in monolayer two-dimensional semiconductors

This paper focuses on the study of linear exciton Hall and Nernst effects in monolayer two-dimensional (2D) semiconductors, employing the semi-classical transport theory. By deriving the exciton Berry curvature in momentum space for a general inhomogeneous 2D system, we establish its dependence on the Berry curvature and the effective mass of electron and hole. As illustrative examples, the exciton Hall effect in monolayer transition metal dichalcogenides (TMDs) and black phosphorus (BP) are calculated. For these materials, we demonstrate that a linear Hall (Nernst) exciton current with the non-zero Berry curvature is strictly forbidden by the symmetries. This finding aligns with earlier experimental observations on the exciton Hall effect in MoSe$_2$. In contrast, a strong anisotropy in BP leads to a net linear Hall current of excitons, exhibiting a relatively large value and resembling an anomalous Hall effect rather than a valley Hall effect. Our work reveals that the specific symmetry of 2D materials can induce a significant linear exciton Hall (Nernst) effect even without Berry curvature, which is normally forbidden with non-zero Berry curvature in the monolayer 2D material. This observation holds promise for future optoelectronic applications and offers exciting possibilities for experimental exploration.

cond-mat.mes-hall↗

arXiv2025

AutoScale: Linear Scalarization Guided by Multi-Task Optimization Metrics

Recent multi-task learning studies suggest that linear scalarization, when using well-chosen fixed task weights, can achieve comparable to or even better performance than complex multi-task optimization (MTO) methods. It remains unclear why certain weights yield optimal performance and how to determine these weights without relying on exhaustive hyperparameter search. This paper establishes a direct connection between linear scalarization and MTO methods, revealing through extensive experiments that well-performing scalarization weights exhibit specific trends in key MTO metrics, such as high gradient magnitude similarity. Building on this insight, we introduce AutoScale, a simple yet effective two-phase framework that uses these MTO metrics to guide weight selection for linear scalarization, without expensive weight search. AutoScale consistently shows superior performance with high efficiency across diverse datasets including a new large-scale benchmark.

cs.LG↗

arXiv2025

Topological classification and edge states of magnons in honeycomb ferromagnets

We study the topological classification and related edge states of magnons in ferromagnets on honeycomb that can be described by a class of single-particle bosonic Bogoliubov-de Gennes (BdG) models. Both single layer and bilayer situations are considered. The calculations show that the existence and related topologies of these edge states are well captured by a class of non-Hermitian single or coupled Su-Schrieffer-Heeger chains models H(ky) parameterized by momentum ky, where the edge states can appear as the ground state for some cases. Interestingly, although the eigenproblem of bosonic BdG models is equivalent to the one of non-Hermitian systems, the conventional bulkedge correspondence for Hermitian systems is partially valid. The influence of Dzyaloshinskii-Moriya interactions between next nearest-neighbor spins are also discussed, which break the time-reversal symmetry and lead to a straight connection between edge states for magnonic systems and non-zero Chern number of non-Hermitian bulk two-dimensional systems.

cond-mat.mes-hall↗

arXiv2024

Dessie: Disentanglement for Articulated 3D Horse Shape and Pose Estimation from Images

In recent years, 3D parametric animal models have been developed to aid in estimating 3D shape and pose from images and video. While progress has been made for humans, it's more challenging for animals due to limited annotated data. To address this, we introduce the first method using synthetic data generation and disentanglement to learn to regress 3D shape and pose. Focusing on horses, we use text-based texture generation and a synthetic data pipeline to create varied shapes, poses, and appearances, learning disentangled spaces. Our method, Dessie, surpasses existing 3D horse reconstruction methods and generalizes to other large animals like zebras, cows, and deer. See the project website at: \url{https://celiali.github.io/Dessie/}.

cs.CV↗

arXiv2024

F\"orster valley-orbit coupling and topological lattice of hybrid moir\'e excitons

Hybrid exciton in moir\'e superlattices of two-dimensional (2D) semiconductors inherits the electric dipole, strong moir\'e trapping, and stacking optical selection rules from its interlayer part, whereas the intralayer part is intended for enhancing optical coupling strength. Here, we show that electron-hole Coulomb exchange, or F\"orster coupling, within the intralayer component qualitatively alters the properties of moir\'e excitons, enabling their coherent hopping between moir\'e traps laterally separated over 10 nm and/or across layers, where their kinetic propagation is completely suppressed. Valley-flip hopping channels are found as significant as the valley-conserving ones, leading to rich possibilities to tailor valley-orbit-couplings and introduce non-trivial topology to the moir\'e exciton superlattice. In twisted MoTe$_2$ where hybrid moir\'e excitons feature a symmetry protection from radiative recombination, we show that F\"orster valley-orbit-coupling can give rise to a rich topological phase diagram.

cond-mat.mes-hall↗

arXiv2024

CLHOP: Combined Audio-Video Learning for Horse 3D Pose and Shape Estimation

In the monocular setting, predicting 3D pose and shape of animals typically relies solely on visual information, which is highly under-constrained. In this work, we explore using audio to enhance 3D shape and motion recovery of horses from monocular video. We test our approach on two datasets: an indoor treadmill dataset for 3D evaluation and an outdoor dataset capturing diverse horse movements, the latter being a contribution to this study. Our results show that incorporating sound with visual data leads to more accurate and robust motion regression. This study is the first to investigate audio's role in 3D animal motion recovery.

cs.CV↗

arXiv2024

Chiral excitonic systems in twisted bilayers from F\"{o}rster coupling and unconventional excitonic Hall effects

In twisted bilayer semiconductors with arbitrary twisting angles, a chiral excitonic system can arise from the interlayer electron-hole Coulomb exchange interaction (F\"{o}rster coupling) that hybridizes the anisotropic intralayer excitons from individual layers. We present a general framework for the effective exciton Hamiltonian taking into account the electron-hole Coulomb exchange, using twisted homobilayer systems composed of transition metal dichalcogenides or black phosphorus as examples. We demonstrate that such chiral excitonic systems can feature unconventional Hall (Nernst) effects arising from quantum geometric properties characteristic of the layer hybridized wavefunctions under the chiral symmetry, for example, the time-reversal even layer Hall counter flow and the crossed nonlinear dynamical Hall effect, when mechanical and statistical force (temperature or density gradient) drives the exciton flow.

cond-mat.mes-hall↗

arXiv2024

Non-Hermitian theory of valley excitons in two-dimensional semiconductors

Electron-hole exchange interaction in two-dimensional transition metal dichalcogenides is extremely strong due to the dimension reduction, which promises valley-superposed excitonic states with linearly polarized optical emissions. However, strong circular polarization reflecting valley-polarized excitonic states is commonly observed in helicity-resolved optical experiments. Here we present a non-Hermitian theory of valley excitons by incorporating optical pumping and intrinsic decay, which unveils an anomalous valley-polarized excitonic state with elliptically polarized optical emission. This novel state arises from the non-Hermiticity induced parity-time ($\mathcal{PT}$)-symmetry breaking, which impedes the experimental observation of intervalley excitonic coherence effect. At large excitonic center-of-mass momenta, the $\mathcal{PT}$-symmetry is restored and the excitonic states recover their valley coherence. Interestingly, the linear polarization directions in optical emissions from these valley-superposed excitonic states are non-orthogonal and even become parallel at exceptional points. Our non-Hermitian theory also predicts a non-zero Berry curvature for valley excitons, which admits a topological excitonic Hall transport beyond the Hermitian predictions.

cond-mat.mes-hall↗

arXiv2023

Human-Centric Autonomous Systems With LLMs for User Command Reasoning

The evolution of autonomous driving has made remarkable advancements in recent years, evolving into a tangible reality. However, a human-centric large-scale adoption hinges on meeting a variety of multifaceted requirements. To ensure that the autonomous system meets the user's intent, it is essential to accurately discern and interpret user commands, especially in complex or emergency situations. To this end, we propose to leverage the reasoning capabilities of Large Language Models (LLMs) to infer system requirements from in-cabin users' commands. Through a series of experiments that include different LLM models and prompt designs, we explore the few-shot multivariate binary classification accuracy of system requirements from natural language textual commands. We confirm the general ability of LLMs to understand and reason about prompts but underline that their effectiveness is conditioned on the quality of both the LLM model and the design of appropriate sequential prompts. Code and models are public with the link \url{https://github.com/KTH-RPL/DriveCmd_LLM}.

cs.CL↗

arXiv2022

Cross-dimensional valley excitons from F\"{o}rster coupling in arbitrarily twisted stacks of monolayer semiconductors

In stacks of transition metal dichalcogenide monolayers with arbitrary twisting angles, we explore a new class of bright excitons arising from the pronounced F\"{o}rster coupling, whose dimensionality is tuned by its in-plane momentum. The low energy sector at small momenta is two-dimensional, featuring a Mexican Hat dispersion, while the high energy sector at larger momenta becomes three-dimensional (3D) with sizable group velocity both in-plane and out-of-plane. By choices of the spacer thickness, interface exciton mode strongly localized at designated layers can emerge out of the cross-dimensional bulk dispersion for a topological origin. Step-edges in spacers can be exploited for engineering lateral interfaces to enable interlayer communication of the topological interface exciton. Combined with the polarization selection rule inherited from the monolayer building block, these exotic exciton properties open up new opportunities for multilayer design towards 3D integration of valley exciton optoelectronics.

cond-mat.mes-hall↗

arXiv2022

Anomalous Bloch oscillation and electrical switching of edge magnetization in bilayer graphene nanoribbon

Graphene features topological edge bands that connect the pair of Dirac points through either sectors of the 1D Brillouin zone depending on edge configurations (zigzag or bearded). Because of their flat dispersion, spontaneous edge magnetisation can arise from Coulomb interaction in graphene nanoribbons, which has caught remarkable interest. We find an anomalous Bloch oscillation in such edge bands, in which the flat dispersion freezes electron motion along the field direction, while the topological connection of the bands through the bulk leads to electron oscillation in the transverse direction between edges of different configurations on opposite sides/layers of a bilayer ribbon. Our Hubbard-model mean-field calculation shows that this phenomenon can be exploited for electrical switching of edge magnetisation configurations.

cond-mat.mes-hall↗

arXiv2021

Edge state in AB-stacked bilayer graphene and its correspondence with SSH ladder

We study edge states in AB-stacked bilayer graphene (BLG) ribbon where the Chern number of the corresponding two-dimensional (2D) bulk Hamiltonian is zero. The existence and topological features of edge states when two layers ended with the same or different edge terminations (zigzag, bearded, armchair) are discussed. The edge states (non-dispersive bands near the Fermi level) are states localized at the edge of graphene nanoribbon that only exists in certain range of momentum $k_y$. Their existence near the Fermi level are protected by the chiral symmetry with topology well described by coupled Su-Schrieffer-Heeger (SSH) chains model, i.e., SSH ladder, based on the bulk-edge correspondence of one-dimensional (1D) systems. These zero-energy edge states can exist in the whole $k_y$ region when two layers have zigzag and bearded edges, respectively. Winding number calculation shows a topological phase transition between two distinct non-trivial topological phases when crossing the Dirac points. Interestingly, we find the stacking configuration of BLG ribbon is important since they can lead to unexpected edge states without protection from the chiral symmetry both near the Fermi level in armchair-armchair case and in the gap within bulk bands that are away from Fermi level in the general case. The influence of interlayer next nearest neighbor (NNN) interaction and interlayer bias are also discussed to fit the realistic graphene materials, which suggest the robust topological features of edge states in BLG systems.

cond-mat.mes-hall↗

arXiv2021

hSMAL: Detailed Horse Shape and Pose Reconstruction for Motion Pattern Recognition

In this paper we present our preliminary work on model-based behavioral analysis of horse motion. Our approach is based on the SMAL model, a 3D articulated statistical model of animal shape. We define a novel SMAL model for horses based on a new template, skeleton and shape space learned from $37$ horse toys. We test the accuracy of our hSMAL model in reconstructing a horse from 3D mocap data and images. We apply the hSMAL model to the problem of lameness detection from video, where we fit the model to images to recover 3D pose and train an ST-GCN network on pose data. A comparison with the same network trained on mocap points illustrates the benefit of our approach.

cs.CV↗

arXiv2021

Revealing the non-adiabatic and non-Abelian multiple-band effects via anisotropic valley Hall conduction in bilayer graphene

Many quantum materials of interest, ex., bilayer graphene, possess a number of closely spaced but not fully degenerate bands near the Fermi level, where the coupling to the far detuned remote bands can induce Berry curvatures of the non-Abelian character in this active multiple-band manifold for transport effects. Under finite electric fields, non-adiabatic interband transition processes are expected to play significant roles in the associated Hall conduction. Here through an exemplified study on the valley Hall conduction in AB-stacked bilayer graphene, we show that the contribution arising from non-adiabatic transitions around the bands near the Fermi energy to the Hall current is not only quantitatively about an order-of-magnitude larger than the contribution due to adiabatic inter-manifold transition with the non-Abelian Berry curvatures. Due to the trigonal warping, the former also displays an anisotropic response to the orientation of the applied electric field that is qualitatively distinct from that of the latter. We further show that these anisotropic responses also reveal the essential differences between the diagonal and off-diagonal elements of the non-Abelian Berry curvature matrix in terms of their contributions to the Hall currents. We provide a physically intuitive understanding of the origin of distinct anisotropic features from different Hall current contributions, in terms of band occupations and interband coherence. This then points to the generalization beyond the specific example of bilayer graphenes.

cond-mat.mes-hall↗

arXiv2020

Theory of wavepacket transport under narrow gaps and spatial textures: non-adiabaticity and semiclassicality

We generalise the celebrated semiclassical wavepacket approach from the adiabatic to the non-adiabatic regime. A unified description covering both of these regimes is particularly desired for systems with spatially varying band structures where band gaps of various sizes are simultaneously present, e.g. in moir\'{e} patterns. For a single wavepacket, alternative to the previous derivation by Lagrangian variational approach, we show that the same semiclassical equations of motion can be obtained by introducing a spatial-texture-induced force operator similar to the Ehrenfest theorem. For semiclassically computing the current, the ensemble of wavepackets based on adiabatic dynamics is shown to well correspond to a phase-space fluid for which the fluid's mass and velocity are two distinguishable properties. This distinction is not inherited to the ensemble of wavepackets with the non-adiabatic dynamics. We extend the adiabatic kinetic theory to the non-adiabatic regime by taking into account decoherence, whose joint action with electric field favours certain form of inter-band coherence. The steady-state density matrix as a function of the phase-space variables is then phenomenologically obtained for calculating the transport current. The result, applicable with a finite electric field, expectedly reproduces the known adiabatic limit by taking the electric field to be infinitesimal, and therefore attains a unified description from the adiabatic to the non-adiabatic situations.

cond-mat.mes-hall↗

arXiv2020

Non-adiabatic Hall effect at Berry curvature hot spot

Hot spot of Berry curvature is usually found at Bloch band anti-crossings, where the Hall effect due to the Berry phase can be most pronounced. With small gaps there, the adiabatic limit for the existing formulations of Hall current can be exceeded in a moderate electric field. Here we present a theory of non-adiabatic Hall effect, capturing non-perturbatively the across gap electron-hole excitations by the electric field. We find a general connection between the field induced electron-hole coherence and intrinsic Hall velocity. In coherent evolution, the electron-hole coherence can manifest as a sizeable ac Hall velocity. When environmental noise is taken into account, its joint action with the electric field favors a form of electron-hole coherence that is function of wavevector and field only, leading to a dc nonlinear Hall effect. The Hall current has all odd order terms in field, and still retains the intrinsic role of the Berry curvature. The quantitative demonstration uses the example of gapped Dirac cones, and our theory can be used to describe the bulk pseudospin Hall current in insulators with gapped edge such as graphene and 2D MnBi$_{2}$Te$_{4}$

cond-mat.mes-hall↗

arXiv2014

Equivalent spin-orbit interaction in two-polariton Jaynes-Cummings-Hubbard model

A hybrid quantum system combines two or more distinct quantum components, exhibiting features not seen in these individual systems. In this work, we study the one-dimensional Jaynes-Cummings- Hubbard model in the two-excitation subspace. We find that the center momentum of two-excitation induces a magnetic ux piercing the 4-leg ladder in the auxiliary space. Furthermore, it is shown that the system in {\pi}-center-momentum subspace is equivalent to a chain system for spin-1 particle with spin-orbit coupling. As a simple application, based on this concise description, a series of bound-pair eigenstates is presented, which displays long-range correlation.

quant-ph↗