arXiv subjects

Ying Zhu

Publications and source records attributed to Ying Zhu.

At least 19 recordsLinked to original sources

arXiv2026

Multiband-Induced Complex Magnetotransport in the Kagome Metal LuNb$_6$Sn$_6$

We report the electrical transport properties of the kagome metal LuNb$_6$Sn$_6$, which undergoes a charge density wave (CDW) transition at $T_\mathrm{CDW}\approx80$ K. Similar to the kagome metals $A$V$_3$Sb$_5$ ($A$ = K, Rb, Cs), LuNb$_6$Sn$_6$ develops complex magnetotransport behavior near and below the CDW transition. In particular, the longitudinal magnetoconductivity exhibits an approximately linear field dependence below about 1 T, whereas the field dependence of Hall conductivity displays pronounced nonlinearity and a field-induced sign reversal. Although an anomalous Hall-like contribution can be extracted by subtracting a local background from the Hall conductivity, mobility-spectrum analysis shows that the overall magnetotransport can be largely accounted for by multiband conduction involving high-mobility electronlike and holelike carriers. Nevertheless, sizable deviations from the multiband model emerge near $T_{\mathrm{CDW}}$, pointing to additional CDW-related scattering processes. These results demonstrate that high-mobility multiband transport, together with CDW-induced modifications of carrier scattering, play important roles in the complex magnetotransport of LuNb$_6$Sn$_6$.

cond-mat.str-el↗

arXiv2026

Emergence of Zonal Flows in an Imbalanced Point-Vortex System

Motivated by experiments on rotating Bose-Einstein condensates, we investigate the statistical mechanics and the non-equilibrium dynamics of an imbalanced gas of oppositely signed Onsager point vortices confined in a disk. To model external forcing and dissipation, we supplement their Hamiltonian dynamics with an annihilation-reinjection (AR) mechanism that preserves the imbalance: oppositely signed vortices are annihilated when they approach closer than a distance representing the healing length, and are subsequently reintroduced into the domain at random locations. At positive energies, the system self-organizes into a large-scale coherent dipole at low imbalance and into a zonal flow at higher imbalance. While AR breaks energy and angular momentum conservation, we argue that the vortex gas evolves quasi-adiabatically through the equilibrium phase space. We capture the topological transition between dipolar and zonal flows both from a statistical and a hydrodynamical perspective. In particular, the bifurcation reflects a non-trivial equilibrium diagram in terms of energy $e$, angular momentum $μ$, and vortex imbalance $η$. We characterize this phase diagram in the "strong mixing'' regime studied by linearizing the Joyce-Montgomery mean-field theory about the weakly non-uniform state. We show that the linear theory is surprisingly relevant for the numerical simulations with AR and suggest that the selection between the dipolar and the zonal configurations in the $(e,μ,η)$ phase space is driven by a hydrodynamical shear instability.

physics.flu-dyn↗

arXiv2026

Hybridization-Induced Enhancement of the Thermoelectric Response in the Kagome Material CsCr$_6$Sb$_6$

Flat bands in kagome lattices provide a prominent route to strongly correlated electronic states. Here we report charge and thermal transport measurements on single crystals of CsCr$_6$Sb$_6$, a kagome Kondo system arising from the hybridization between flat and dispersive bands. The Seebeck coefficient and Nernst signal increase rapidly below $T_\mathrm{F}\sim70$ K, concomitant with the emergence of short-range antiferromagnetic correlations and flat-band resonance. The Nernst coefficient in the intermediate-temperature range of $\sim30-80$ K further follows the universal scaling reported for prototypical heavy-fermion compounds, consistent with the development of strongly renormalized quasiparticles associated with hybridization. Both the Seebeck and Nernst responses reach extrema at a lower temperature $T_\mathrm{p}\sim20-25$ K, where phonon drag and correlated quasiparticles may both contribute. These results reveal a hybridization-induced enhancement of the thermoelectric response and establish CsCr$_6$Sb$_6$ as a kagome platform for probing flat-band-derived Kondo physics through thermoelectric transport.

cond-mat.str-el↗

arXiv2026

Relaxation Dynamics of Superconducting/Normal Interfaces in the Intermediate State of a type-I superconductor

The dynamics of phase boundaries, such as superconducting/normal (S/N) interfaces in type-I superconductors, are typically obscured in conventional magnetic measurements, which are strongly influenced by surface barriers and global flux processes. Here, we employ the complex ac magnetostrictive coefficient as a sensitive probe of the bulk dynamics of S/N interfaces in the intermediate state of lead. The ac magnetostrictive coefficient exhibits an unusual complex response, with a negative real component and a positive imaginary component. Frequency-dependent measurements show that the magnitude of the real component decreases monotonically with increasing frequency, whereas the imaginary component first increases, reaches a broad maximum, and then decreases. In contrast, ac susceptibility displays a conventional Debye-type magnetic relaxation with a different phase relation between its real and imaginary components. We propose that the magnetostrictive response arises from ac-field-driven relaxation and rearrangement of metastable S/N domain configurations toward states of lower free energy. These results reveal a distinct interfacial relaxation channel in the intermediate state and establish the complex ac magnetostrictive coefficient as a sensitive probe of nonequilibrium S/N interface dynamics.

cond-mat.supr-con↗

arXiv2026

Complementing reinforcement learning with SFT through logit averaging in the post training of LLMs

We introduce a novel method that averages the logits of a frozen reference policy (e.g., SFT) and a trainable policy, and incorporate the method into Group Relative Policy Optimization (GRPO). In contrast to Reinforcement Learning with Verifiable Rewards (RLVR) methods, our proposal does not involve a Kullback Leibler (KL) regularization or critic; the trainable policy and the reference anchor are coupled through the logit averaging structure to leverage the reasoning expertise of the trainable policy while maintaining the formatting advantage of SFT. Our method is evaluated on MATH, cn-k12, and MMLU, and the results show a higher accuracy or at least comparable accuracy relative to the canonical KL-regularized GRPO.

cs.LG↗

arXiv2026

Phase calibration of quantum oscillations in the magnetostrictive coefficient using the topological antiferromagnet YbMnBi$_2$

The Berry phase accumulated along a cyclotron orbit encodes important information about electronic band topology and is commonly inferred from the phase of quantum oscillations. Measurements of the ac magnetostrictive coefficient have recently emerged as a sensitive thermodynamic probe of quantum oscillations, but the phase offset has not been experimentally calibrated. Here, using the topological antiferromagnet YbMnBi$_2$, we calibrate this offset by directly comparing quantum oscillations in magnetization with those in the ac magnetostrictive coefficient. Measurements of both responses on the same single crystal reveal a single fundamental frequency of approximately 160 T in fields up to 14 T, enabling a direct phase comparison free from ambiguities associated with multiple frequencies. We observe an approximately $π/2$ relative phase shift between the two oscillatory responses, consistent with the Maxwell relation linking the magnetostrictive coefficient to the stress derivative of magnetization. Our results establish the appropriate phase needed to extract cyclotron-orbit phase information from quantum oscillations in the ac magnetostrictive coefficient.

cond-mat.str-el↗

arXiv2026

Approximating invariant functions with the sorting trick is theoretically justified

Many machine learning models leverage group invariance which is enjoyed with a wide-range of applications. For exploiting an invariance structure, one common approach is known as \emph{frame averaging}. One popular example of frame averaging is the \emph{group averaging}, where the entire group is used to symmetrize a function. Another example is the \emph{canonicalization}, where a frame at each point consists of a single group element which transforms the point to its orbit representative, for example, sorting. Compared to group averaging, canonicalization is more efficient computationally. However, it results in non-differentiability or discontinuity of the canonicalized function. As a result, the theoretical performance of canonicalization has not been given much attention. In this work, we establish an approximation theory for canonicalization. Specifically, we bound the point-wise and $L^2(\mathbb{P})$ approximation errors as well as the eigenvalue decay rates associated with a canonicalization trick applied to reproducing kernels. We discuss two key insights from our theoretical analyses and why they point to an interesting future research direction on how one can choose a design to fully leverage canonicalization in practice.

cs.LG↗

arXiv2026

FourierSampler: Unlocking Non-Autoregressive Potential in Diffusion Language Models via Frequency-Guided Generation

Despite the non-autoregressive potential of diffusion language models (dLLMs), existing decoding strategies demonstrate positional bias, failing to fully unlock the potential of arbitrary generation. In this work, we delve into the inherent spectral characteristics of dLLMs and present the first frequency-domain analysis showing that low-frequency components in hidden states primarily encode global structural information and long-range dependencies, while high-frequency components are responsible for characterizing local details. Based on this observation, we propose FourierSampler, which leverages a frequency-domain sliding window mechanism to dynamically guide the model to achieve a "structure-to-detail" generation. FourierSampler outperforms other inference enhancement strategies on LLADA and SDAR, achieving relative improvements of 20.4% on LLaDA1.5-8B and 16.0% on LLaDA-8B-Instruct. It notably surpasses similarly sized autoregressive models like Llama3.1-8B-Instruct.

cs.CL↗

arXiv2026

OmniVAE: An Audio-Video VAE with Cross-Modal Alignment for Joint Generation

Recent generative models are moving beyond silent video or standalone audio synthesis toward the joint generation of synchronized audio and video. Despite this progress, jointly generating audio and video with fine-grained cross-modal correspondence remains challenging due to their fundamental structural differences. Most existing methods use audio and video VAEs trained separately. As a result, the two latent spaces lack cross-modal alignment, leaving the downstream generative model to learn cross-modal synchronization from scratch. We present OmniVAE, a jointly trained audio-video VAE that learns fine-grained semantic alignment between audio and video latent representations. Beyond reconstruction, OmniVAE uses a segment-level audio-video contrastive objective to capture temporal-semantic correspondence and align the two latent spaces. In parallel, it distills features from pretrained modality-specific semantic encoders into each modality, improving the downstream learnability of both latent spaces. Extensive experiments show that both objectives consistently improve the learnability of the latent spaces, translating into higher generation quality and more accurate cross-modal synchronization in downstream text-to-audio-video generation. These findings underscore the importance of learning unified representations as a foundation for omnimodal modeling.1

cs.SD↗

arXiv2026

Complex Temperature-dependent Thermal Conductivity in a Sawtooth Chain Magnet Fe$_\mathrm{2}$SiSe$_\mathrm{4}$

Geometrically frustrated magnets provide an ideal platform for exploring the interplay between lattice geometry and spin degrees of freedom. Here, we investigate the interactions between lattice and spin via thermal-transport measurements on the triangular sawtooth-lattice olivine magnet Fe$_\mathrm{2}$SiSe$_\mathrm{4}$, which exhibits successive magnetic transitions at $T_1 = 110$ K (antiferromagnetic) and $T_2 = 50$ K (ferrimagnetic). Although phonons dominate the thermal conductivity, its temperature dependence displays a pronounced double-peak structure arising from spin-phonon coupling. In the intermediate temperature range between $T_1$ and $T_2$ , resonant scattering of phonons by magnetic excitations around 5 meV produces a broad maximum around 60 K. Below $T_2$, the resonant spin-phonon scattering is strongly suppressed, leading to a rapid increase in thermal conductivity upon cooling and a pronounced low-temperature peak near 11 K, characteristic of heat transport governed by conventional phonon scattering mechanisms. Notably, this low-temperature peak is enhanced by a factor of $\sim 5$ compared to the broad maximum at higher temperatures. These results demonstrate the strong sensitivity of thermal transport to spin-lattice interactions and highlight spin-phonon scattering as an effective mechanism for tailoring thermal conductivity in geometrically frustrated magnets.

cond-mat.str-el↗

arXiv2026

Turbulence and far-from-equilibrium equation of state of Bogoliubov waves in Bose-Einstein Condensates

Bogoliubov waves are fundamental excitations of Bose-Einstein Condensates (BECs). They emerge from a perturbed ground state and interact nonlinearly, triggering turbulent cascades. Here, we study turbulent BECs theoretically and numerically using the 3D Gross-Pitaevskii model and its associated wave-kinetic equations. We derive a new Kolmogorov-like stationary spectrum for short Bogoliubov waves and find a complete analytical expression for the spectrum in the long-wave acoustic regime. We then use our predictions to explain the BEC equation of state reported by Dogra et al. (Nature 620, 521, 2023), and to suggest new experimental settings.

cond-mat.quant-gas↗

arXiv2026

Less is More: Lightweight Prompt Compression for Question Answering Applications on Edge Devices

In agent-driven question answering (QA) applications, retrieval-augmented generation (RAG) is commonly introduced to enhance the response accuracy of large language models (LLMs) by providing additional context. Due to the inherent noise in retrieval results and the coarse granularity of document-level retrieval, the retrieved context often contains substantial redundant information. In this setting, the agent prompt, consisting of the user query and the associated retrieved context, leads to unnecessary computational overhead during LLM inference. Existing prompt compression methods typically rely on auxiliary small language models (SLMs) to estimate context importance. However, such approaches introduce significant memory and computational overhead, which limits their deployment on resource-constrained edge devices. In this paper, we propose CORE, a two-stage sentence-level prompt compression method that eliminates the need for SLMs. In the first stage, CORE constructs an answer set via named entity recognition (NER) and a clue set via semantic matching. In the second stage, CORE refines the clue set using an orthogonal residual retrieval strategy and designs a spatial proximity-based metric to filter the answer set. The two sets are then combined to form the final compressed context. We implement CORE on an NVIDIA Jetson AGX Orin edge device and a Huawei Nova smartphone. Experimental results demonstrate that within a 2000-token budget, CORE improves accuracy by at least 30.19% compared to state-of-the-art baselines, while reducing memory usage by at least 50.47% and achieving at least 1.94 times speedup on the edge device. Moreover, compared to the state-of-the-art LLMLingua2 method, CORE achieves a substantial energy reduction of 95.74% on the smartphone, highlighting its practicality and generalizability for mobile deployments.

cs.CL↗

arXiv2026

Strong and weak wave turbulence regimes in Bose-Einstein condensates

When a turbulent Bose-Einstein condensate is driven out-of-equilibrium at a scale much smaller than the system size, nonlinear wave interactions transfer particles towards large scales in an inverse cascade process. In this work, we study numerically wave turbulence in a three-dimensional Bose-Einstein condensate in forced and dissipated inverse cascade settings. We observe that when the forcing rate increases, thereby increasing the particle flux, the turbulence spectrum gradually transitions from the weak-wave Kolmogorov-Zakharov cascade to a critical balance state characterized by a range of scales with balanced linear and nonlinear dynamic timescales. Further forcing increases lead to a coherent condensate component superimposed with Bogoliubov-type acoustic turbulence. The role of vortices in such a strongly forced state is marginal, which makes this new state very different from the strongly turbulent state composed of a tangle of quantized vortex lines. We then use our predictions and numerical data to formulate a new out-of-equilibrium equation of state for the 3D inverse cascade.

cond-mat.quant-gas↗

arXiv2026

Evaluating Prompting Strategies for Chart Question Answering with Large Language Models

Prompting strategies affect LLM reasoning performance, but their role in chart-based QA remains underexplored. We present a systematic evaluation of four widely used prompting paradigms (Zero-Shot, Few-Shot, Zero-Shot Chain-of-Thought, and Few-Shot Chain-of-Thought) across GPT-3.5, GPT-4, and GPT-4o on the ChartQA dataset. Our framework operates exclusively on structured chart data, isolating prompt structure as the only experimental variable, and evaluates performance using two metrics: Accuracy and Exact Match. Results from 1,200 diverse ChartQA samples show that Few-Shot Chain-of-Thought prompting consistently yields the highest accuracy (up to 78.2\%), particularly on reasoning-intensive questions, while Few-Shot prompting improves format adherence. Zero-Shot performs well only with high-capacity models on simpler tasks. These findings provide actionable guidance for selecting prompting strategies in structured data reasoning tasks, with implications for both efficiency and accuracy in real-world applications.

cs.CL↗

arXiv2026

Universal Behavior on the Relaxation Dynamics of Far-From-Equilibrium Quantum Fluids

Investigating the initial conditions that lead many-body quantum systems to an out-of-equilibrium state is fundamental for understanding their thermalization dynamics. In this work we observe the relaxation for two regimes of excitation that can drive the turbulent Bose-Einstein condensate into two distinct final states, and are defined by the amount of energy injected into the system. The subcritical regime is characterized by a lower injection of energy, which can lead to an inverse particle cascade and, consequently, to the BEC mode repopulation during the relaxation process. The supercritical regime is marked by a higher energy injection, that may lead to the BEC dissolution and a final thermal state. In both cases we observe relaxation stages that exhibit the same key features: a direct cascade, a non-thermal fixed point with the same exponents, a prethermalization region and, finally, the thermalization of the system. In the final thermalization stage, universal scaling is observed for both regimes, even though their final states are completely different. By analyzing the coherence length of our turbulent cloud, we clearly visualize the recovery and the loss of the coherence for the subcritical and supercritical regimes after relaxation. These results indicate that the evolution of turbulence occurs independent of its initial conditions and of the final state achieved.

cond-mat.quant-gas↗

arXiv2026

Predicting Tennis Serve directions with Machine Learning

Serves, especially first serves, are very important in professional tennis. Servers choose their serve directions strategically to maximize their winning chances while trying to be unpredictable. On the other hand, returners try to predict serve directions to make good returns. The mind game between servers and returners is an important part of decision-making in professional tennis matches. To help understand the players' serve decisions, we have developed a machine learning method for predicting professional tennis players' first serve directions. Through feature engineering, our method achieves an average prediction accuracy of around 49\% for male players and 44\% for female players. Our analysis provides some evidence that top professional players use a mixed-strategy model in serving decisions and that fatigue might be a factor in choosing serve directions. Our analysis also suggests that contextual information is perhaps more important for returners' anticipatory reactions than previously thought.

cs.LG↗

arXiv2026

MOVA: Towards Scalable and Synchronized Video-Audio Generation

Audio is indispensable for real-world video, yet generation models have largely overlooked audio components. Current approaches to producing audio-visual content often rely on cascaded pipelines, which increase cost, accumulate errors, and degrade overall quality. While systems such as Veo 3 and Sora 2 emphasize the value of simultaneous generation, joint multimodal modeling introduces unique challenges in architecture, data, and training. Moreover, the closed-source nature of existing systems limits progress in the field. In this work, we introduce MOVA (MOSS Video and Audio), an open-source model capable of generating high-quality, synchronized audio-visual content, including realistic lip-synced speech, environment-aware sound effects, and content-aligned music. MOVA employs a Mixture-of-Experts (MoE) architecture, with a total of 32B parameters, of which 18B are active during inference. It supports IT2VA (Image-Text to Video-Audio) generation task. By releasing the model weights and code, we aim to advance research and foster a vibrant community of creators. The released codebase features comprehensive support for efficient inference, LoRA fine-tuning, and prompt enhancement.

cs.CV↗

arXiv2026

Controllable Dance Generation with Style-Guided Motion Diffusion

Dance plays an important role as an artistic form and expression in human culture, yet automatically generating dance sequences is a significant yet challenging endeavor. Existing approaches often neglect the critical aspect of controllability in dance generation. Additionally, they inadequately model the nuanced impact of music styles, resulting in dances that lack alignment with the expressive characteristics inherent in the conditioned music. To address this gap, we propose Style-Guided Motion Diffusion (SGMD), which integrates the Transformer-based architecture with a Style Modulation module. By incorporating music features with user-provided style prompts, the SGMD ensures that the generated dances not only match the musical content but also reflect the desired stylistic characteristics. To enable flexible control over the generated dances, we introduce a spatial-temporal masking mechanism. As controllable dance generation has not been fully studied, we construct corresponding experimental setups and benchmarks for tasks such as trajectory-based dance generation, dance in-betweening, and dance inpainting. Extensive experiments demonstrate that our approach can generate realistic and stylistically consistent dances, while also empowering users to create dances tailored to diverse artistic and practical needs. Code is available on Github: https://github.com/mucunzhuzhu/DGSDP

cs.CV↗