arXiv subjects

Chen Fang

Publications and source records attributed to Chen Fang.

At least 19 recordsLinked to original sources

arXiv2026

Foresight: Learning Twisted SMC for Rare Events from Reversed Relaxations

We consider the probability that a reversible Markov chain started in a set $A$ is in a rare set $B$ after a fixed horizon, a quantity often needed for many start laws, down to single states. Twisted sequential Monte Carlo (SMC) estimates such probabilities without bias, with small variance when its proposal and twist approximate the dynamics conditioned on reaching $B$, but these components are usually learned from forward simulations, which contain almost no successful paths when the event is rare. We introduce Foresight-SMC, which learns both components from the target side: under reversibility, relaxations started in $B$ that return to $A$ are, read backwards, exact samples of the conditioned paths, and their number is governed by the return probability rather than by the rare probability. Extending this reversal to the recorded propose--accept decisions of a Metropolis-adjusted sampler, we fit a controlled Langevin proposal with exact likelihood ratios by maximum likelihood, together with a classifier twist; the resulting estimator is unbiased for any start law. A variance analysis shows that resampling reduces the variance when the importance weights degenerate, as they do over long horizons. On an Allen--Cahn field, the learned proposal alone suffices, and one trained model is more accurate than splitting at equal cost. On alanine tripeptide, resampling with the learned twist is necessary, and Foresight-SMC needs less computation per query than adaptive multilevel splitting on localized start sets.

cs.LG↗

arXiv2026

Crystal symmetry predicts unconventional magnetism

Unconventional compensated magnets combine zero net magnetization with momentum-dependent spin polarization, but identifying them usually requires knowledge of their magnetic order. Here we show that crystal symmetry can constrain unconventional magnetic character before the magnetic ground state is known. Starting from a non-magnetic crystal structure and a specified magnetic sublattice, we generate symmetry-compatible compensated orders and classify their spin textures using spin-space-group symmetry. We identify materials whose generated candidates are all unconventional, either across a defined search space or after restricting the magnetic-cell size. Within the experimental benchmark, 68% of the prioritized materials are unconventional, compared with 9% of the remaining materials. Screening the Materials Project yields thousands of promising candidates for unconventional compensated magnetism. First-principles calculations for $\mathrm{VGe_3}$ and tetragonal $\mathrm{Fe_2SiO_4}$ connect these symmetry predictions to the energetics and spin textures of competing magnetic orders. In $\mathrm{VGe_3}$, a noncoplanar candidate permits mixed-wave spin polarization along a fixed axis without spin-orbit coupling, combining components that are odd and even under momentum reversal. This framework enables crystallography-guided searches for unconventional compensated magnets without first determining their magnetic ground states.

cond-mat.mtrl-sci↗

arXiv2026

General Construction from Topological Loop States to Topological Invariants in Exactly Flat Bands

Electronic flat bands have localized Wannier-like orbitals as zero modes. In the Lieb or the kagome models, the localized orbitals satisfy a topological condition that entails two non-contractible loop eigenstates along $x/y$-axis in real space, and one topological band touching point with other bands in momentum space. In these topological-flat bands, the Bloch state at the touching point is ill-defined, and so is any topological invariant for the entire band. We propose a new topological condition that the loop states in different directions be linearly dependent. Its satisfaction removes the singularity at the band touching point, and enforces nontrivial, well-defined topological invariants. Enforcing the new condition, we obtain topological-topological (top$^2$)-flat bands in 2D and 3D that have nontrivial invariants including the Chern numbers, the $\mathbb{Z}_2$ invariants, and the topological-crystalline invariants. Under small, generic interactions, top$^2$-flat bands flow to correlated topological insulators with a symmetric mass term; and specially designed interacting models can have top$^2$-flat bands as exact zero modes.

cond-mat.mes-hall↗

arXiv2026

Algebraic States in Continuum in $ d\gt 1$ Dimensional Non-Hermitian Systems

We report the existence of algebraically localized eigenstates embedded within the continuum spectrum of 2D non-Hermitian systems with a single impurity. These modes, which we term algebraic states in continuum (AICs), decay algebraically as $1/|r|$ from the impurity site, and their energies lie within the bulk continuum spectrum under periodic boundary conditions. We analytically derive the threshold condition for the impurity strength required to generate such states. Remarkably, AICs are forbidden in Hermitian systems and in 1D non-Hermitian systems, making them unique to non-Hermitian systems in two and higher dimensions. To detect AICs, we introduce a local density of states as an experimental observable, which is readily accessible in photonic/acoustic platforms.

cond-mat.mes-hall↗

arXiv2026

Dynamical Poles in Non-Hermitian Impurity Scattering

In Hermitian impurity scattering, each isolated late-time exponential is the fingerprint of a bound state. We show that this correspondence breaks down in non-Hermitian bands. For a single impurity in a non-Hermitian lattice, the late-time signal is controlled by isolated complex frequencies selected by the analytic continuation of the Green's function relevant to real-time dynamics, which we term dynamical poles (DPs). DPs need not coincide with static bound states: one may appear without any bound-state counterpart, while a static bound state may be dynamically invisible. The remainder of the signal is an incoherent background set by complex continuum edges. Our results establish that the real-time analytic structure of the Green's function, not the static eigenvalue problem alone, organizes non-Hermitian impurity scattering.

cond-mat.mes-hall↗

arXiv2026

AudioGuard: Toward Comprehensive Audio Safety Protection Across Diverse Threat Models

Audio has rapidly become a primary interface for foundation models, powering real-time voice assistants. Ensuring safety in audio systems is inherently more complex than just "unsafe text spoken aloud": real-world risks can hinge on audio-native harmful sound events, speaker attributes (e.g., child voice), impersonation/voice-cloning misuse, and voice-content compositional harms, such as child voice plus sexual content. The nature of audio makes it challenging to develop comprehensive benchmarks or guardrails against this unique risk landscape. To close this gap, we conduct large-scale red teaming on audio systems, systematically uncover vulnerabilities in audio, and develop a comprehensive, policy-grounded audio risk taxonomy and AudioSafetyBench, the first policy-based audio safety benchmark across diverse threat models. AudioSafetyBench supports diverse languages, suspicious voices (e.g., celebrity/impersonation and child voice), risky voice-content combinations, and non-speech sound events. To defend against these threats, we propose AudioGuard, a unified guardrail consisting of 1) SoundGuard for waveform-level audio-native detection and 2) ContentGuard for policy-grounded semantic protection. Extensive experiments on AudioSafetyBench and four complementary benchmarks show that AudioGuard consistently improves guardrail accuracy over strong audio-LLM-based baselines with substantially lower latency.

cs.SD↗

arXiv2026

Deep Generative Models in Condition and Structural Health Monitoring: Opportunities, Limitations and Future Outlook

Condition and structural health monitoring (CM/SHM) is a pivotal component of predictive maintenance (PdM) strategies across diverse industrial sectors, including mechanical rotating machinery, aircraft structures, wind turbines, and civil infrastructures. Conventional deep learning models, while effective for fault diagnosis and anomaly detection through automatic feature learning from sensor data, often struggle with operational variability, imbalanced or scarce fault datasets, and multimodal sensory data from complex systems. Deep generative models (DGMs) including deep autoregressive models, variational autoencoders, generative adversarial networks, diffusion-based models, and emerging large language models, offer transformative capabilities by synthesizing high-fidelity data samples, reconstructing latent system states, and modeling complex multimodal data streams. This review systematically examines state-of-the-art DGM applications in CM/SHM across the four main industrial systems mentioned above, emphasizing their roles in addressing key challenges: data generation, domain adaptation and generalization, multimodal data fusion, and downstream fault diagnosis and anomaly detection tasks, with rigorous comparison among signal processing, conventional machine learning or deep learning models, and DGMs. Lastly, we discuss current limitations of DGMs, including challenges of explainable and trustworthy models, computational inefficiencies for edge deployment, and the need for parameter-efficient fine-tuning strategies. Future research directions can focus on zero-shot and few-shot learning, robust multimodal data generation, hybrid architectures integrating DGMs with physics knowledge, and reinforcement learning with DGMs to enhance robustness and accuracy in industrial scenarios.

cs.CE↗

arXiv2026

Tailoring Bound State Geometry in High-Dimensional Non-Hermitian Systems

It is generally believed that the non-Hermitian effect (NHSE), due to its non-reciprocal nature, creates barriers for the appearance of impurity bound states. In this paper, we find that in two and higher dimensions, the presence of geometry-dependent skin effect eliminates this barrier such that even an infinitesimal impurity potential can confine bound states in this type of non-Hermitian systems. By examining bound states around Bloch saddle points, we find that non-Hermiticity can disrupt the isotropy of bound states, resulting in concave dumbbell-shaped bound states. Our work reveals a geometry transition of bound state between concavity and convexity in high-dimensional non-Hermitian systems, offering theoretical insights for the experimental manipulation of bound states.

cond-mat.mes-hall↗

arXiv2026

Resolution of Topology and Geometry from Momentum-Resolved Spectroscopies

Extracting the complete quantum geometric and topological character of Bloch wavefunctions from experiments remains a challenge in condensed matter physics. Here, we resolve this by introducing the "wavefunction form factor" (WFF) matrix, a quantity directly constructible from intensities in momentum- and energy-resolved spectroscopies like ARPES and INS. We demonstrate that band topology is encoded in "spectral nodes" -- momentum-space points where the WFF determinant vanishes, providing a direct readout of topological invariants via a topological selection rule. Furthermore, when the number of independent probes exceeds the number of the target bands, our framework yields an effective band projector. This enables the determination of Wilson loop spectra and the extraction of an effective quantum geometric tensor, providing a model-independent measurement of the non-Abelian Berry curvature and quantum metric as resolved by the experimental probes.

cond-mat.mes-hall↗

arXiv2025

A Unified Symmetry Classification of Magnetic Orders via Spin Space Groups: Prediction of Coplanar Even-Wave Phases

Spin space groups (SSGs) impose fundamentally different constraints on magnetic configurations in real and reciprocal spaces. As a consequence, the correspondence between real-space and momentum-space spin arrangements is far richer than traditionally assumed. Building on the complete enumeration of SSGs, we develop a systematic, symmetry-based framework that classifies all possible spin arrangements allowed by these groups. This unified approach naturally incorporates conventional magnetic orders, altermagnetism, and p-wave magnetism as distinct symmetry classes. Crucially, our classification predicts a variety of novel magnetic phases, highlighted by the discovery of the coplanar even-wave magnet: a state that is non-collinear in real space but hosts a collinear even-wave spin polarization in k-space. Analysis of a minimal model reveals that this phase is characterized by non-quantized spin polarization and exhibits a novel mechanism for symmetry-enforced zero polarization on non-degenerate bands. Extending the framework from bulk crystals to layer SSGs appropriate for two-dimensional systems, we further predict layered counterparts and provide symmetry guidelines for designing bilayer coplanar p-wave and even-wave magnets. We further validate this finding through first-principles calculations and propose CoCrO4 as a promising candidate for its experimental realization, thereby demonstrating the completeness and predictive power of the SSG-based classification of magnetic orders.

cond-mat.mtrl-sci↗

arXiv2025

PersonaLive! Expressive Portrait Image Animation for Live Streaming

Current diffusion-based portrait animation models predominantly focus on enhancing visual quality and expression realism, while overlooking generation latency and real-time performance, which restricts their application range in the live streaming scenario. We propose PersonaLive, a novel diffusion-based framework towards streaming real-time portrait animation with multi-stage training recipes. Specifically, we first adopt hybrid implicit signals, namely implicit facial representations and 3D implicit keypoints, to achieve expressive image-level motion control. Then, a fewer-step appearance distillation strategy is proposed to eliminate appearance redundancy in the denoising process, greatly improving inference efficiency. Finally, we introduce an autoregressive micro-chunk streaming generation paradigm equipped with a sliding training strategy and a historical keyframe mechanism to enable low-latency and stable long-term video generation. Extensive experiments demonstrate that PersonaLive achieves state-of-the-art performance with up to 7-22x speedup over prior diffusion-based portrait animation models.

cs.CV↗

arXiv2025

StereoWorld: Geometry-Aware Monocular-to-Stereo Video Generation

The growing adoption of XR devices has fueled strong demand for high-quality stereo video, yet its production remains costly and artifact-prone. To address this challenge, we present StereoWorld, an end-to-end framework that repurposes a pretrained video generator for high-fidelity monocular-to-stereo video generation. Our framework jointly conditions the model on the monocular video input while explicitly supervising the generation with a geometry-aware regularization to ensure 3D structural fidelity. A spatio-temporal tiling scheme is further integrated to enable efficient, high-resolution synthesis. To enable large-scale training and evaluation, we curate a high-definition stereo video dataset containing over 11M frames aligned to natural human interpupillary distance (IPD). Extensive experiments demonstrate that StereoWorld substantially outperforms prior methods, generating stereo videos with superior visual fidelity and geometric consistency. The project webpage is available at https://ke-xing.github.io/StereoWorld/.

cs.CV↗

arXiv2025

Exceptional points and spectral cusps from density-wave fluctuation

We report two types of singularities that arise from fluctuations during the formation of charge- or spin-density waves. The first is the exceptional point (EP), corresponding to a higher-order pole of the retarded Green's function. Such EPs lead to algebraic corrections in the decay of quasiparticle occupations and are observable through time-resolved angle-resolved photoemission spectroscopy (Tr-ARPES). The second is a spectral cusp, defined by the coalescence of three extrema in the real-frequency spectral function $A(\mathbf{k}, ω)$. This cusp enforces the formation of Fermi arcs and induces a "threading" structure in the nearby band structure, both of which are directly observable in ARPES.

cond-mat.mes-hall↗

arXiv2025

Superdiffusive transport protected by topology and symmetry in all dimensions

Superdiffusion is an anomalous transport behavior. Recently, a new mechanism, termed the ``nodal mechanism," has been proposed to induce superdiffusion in quantum models. However, existing realizations of the nodal mechanism have so far been proposed on fine-tuned, artificial Hamiltonians, posing a significant challenge for experimental observation. In this work, we propose a broad class of models for generating superdiffusion potentially realizable in condensed matter systems across different spatial dimensions. A robust nodal structure emerges from the hybridization between the itinerant electrons and the local impurity orbitals, protected by the intrinsic symmetry and topology of the electronic band. We derive a universal scaling law for the conductance, $G \sim L^{-γ}$, revealing how the exponent is dictated by the dimensionality of the nodal structure ($D_{\text{node}}$) and its order $n$, and the dimensionality of the system $(D)$ at high temperatures or that of the Fermi surface ($D^F$) at low temperatures. Through numerical simulations, we validate these scaling relations at zero temperature for various models, including those based on graphene and multi-Weyl semimetals, finding excellent agreement between our theory and the computed exponents. Beyond the scaling of conductance, our framework predicts a suite of experimentally verifiable signatures, notably a new mechanism for linear-in-temperature resistivity ($ρ\sim T$) and a divergent low-frequency optical conductivity ($σ(ω) \sim ω^{γ-1}$), establishing a practical route to discovering and engineering anomalous transport in quantum materials.

cond-mat.mes-hall↗

arXiv2025

Ultra High-Resolution Image Inpainting with Patch-Based Content Consistency Adapter

In this work, we present Patch-Adapter, an effective framework for high-resolution text-guided image inpainting. Unlike existing methods limited to lower resolutions, our approach achieves 4K+ resolution while maintaining precise content consistency and prompt alignment, two critical challenges in image inpainting that intensify with increasing resolution and texture complexity. Patch-Adapter leverages a two-stage adapter architecture to scale the diffusion model's resolution from 1K to 4K+ without requiring structural overhauls: (1) Dual Context Adapter learns coherence between masked and unmasked regions at reduced resolutions to establish global structural consistency; and (2) Reference Patch Adapter implements a patch-level attention mechanism for full-resolution inpainting, preserving local detail fidelity through adaptive feature fusion. This dual-stage architecture uniquely addresses the scalability gap in high-resolution inpainting by decoupling global semantics from localized refinement. Experiments demonstrate that Patch-Adapter not only resolves artifacts common in large-scale inpainting but also achieves state-of-the-art performance on the OpenImages and Photo-Concept-Bucket datasets, outperforming existing methods in both perceptual quality and text-prompt adherence.

cs.CV↗

arXiv2025

Real-time edge dynamics of non-Hermitian lattices

We derive the asymptotic forms of the Green's function at the open edges of general non-Hermitian band systems in all dimensions in the long-time limit, using a modified saddle-point approximation and the analytic continuation of the momentum. The edge dynamics is determined by the "dominant saddle point", a complex momentum, which, contrary to previous conjectures, may lie outside the generalized Brillouin zone. From this result, we obtain the effective edge Hamiltonians that evidently, as demonstrated by extensive numerical simulations, characterize the dynamics on the edges, and can be probed in real-time experiments or spectroscopies.

cond-mat.mes-hall↗

arXiv2025

Soliton Transitions Mediated by Skin-Mode Localization and Band Nonreciprocity

Solitons, typically resulting from a competition between band dispersion and nonlinearity, occur in lattices featuring the non-Hermitian skin effect as nonlinearity increases, accompanied by a transition in localization from linear skin modes to solitons. However, localization does not disentangle the role of skin modes in the soliton formation from that of band dispersion. Here, in such lattices, we uncover two distinct soliton phases, skin-mode-assisted solitons (SMASs) and nonreciprocity-dressed solitons (NRDSs). Rooted in fundamentally different mechanisms, SMASs originate from skin effect, while NRDSs stem from band nonreciprocity, each exhibiting unique spatial profiles. Using a stacked Su-Schrieffer-Heeger-like model as a prototype, we delineate the phase diagram of SMASs and NRDSs, each having clear phase boundaries. To interpret them, we formulate a Wannier-function-based nonlinear Hamiltonian, showing that soliton formation depends critically on how skin-mode localization and band nonreciprocity suppress or enhance wave dispersion. For SMASs, skin-mode localization reduces wave broadening at the localization sites, thereby lowering the formation threshold. This soliton phase is observable from edge dynamics and accompanied by a dynamical stability reentrance when transitioning from linear skin modes. In contrast, NRDSs, as well as their thresholds, originate from bulk band nonreciprocity and persist under periodic boundary conditions. Our framework offers predictive tools for characterizing and engineering solitons in experimentally realizable non-Hermitian systems, spanning optics to mechanics.

nlin.PS↗

arXiv2025

Baseline-free Damage Detection and Localization on Composite Structures with Unsupervised Kolmogorov-Arnold Autoencoder and Guided Waves

Structural health monitoring (SHM) ensures the safety and longevity of structures such as aerospace equipment and wind power installations. Developing a simple, highly flexible, and scalable SHM method that does not depend on baseline models is significant for ensuring the operational integrity of advanced composite structures. In this regard, a hybrid baseline-free damage detection and localization framework incorporating an unsupervised Kolmogorov-Arnold autoencoder (KAE) and modified probabilistic elliptical imaging algorithm (MRAPID) is proposed for damage detection and localization in composite structures. Specifically, KAE was used to process the guided wave signals (GW) without any prior feature extraction process. The KAE continuously learns and adapts to the baseline model of each structure, learning from the response characteristics of its undamaged state. Then, the predictions from KAE are processed, combined with the MRAPID to generate a damage probability map. The performance of the proposed method for damage detection and localization was verified using the simulated damage data obtained on wind turbine blades and the actual damage data obtained on composite flat plates. The results show that the proposed method can effectively detect and localize damage and can achieve multiple damage localization. In addition, the method outperforms classical damage detection algorithms and state-of-the-art baseline-free damage detection and localization methods in terms of damage localization accuracy.

cs.CE↗