Searcharxiv⌕ Search

arXiv subjects

Zhifei Zhang

Publications and source records attributed to Zhifei Zhang.

At least 19 recordsLinked to original sources

On the absence of point defects in biaxial Landau--de Gennes models

We study local minimizers of a sextic-potential Landau--de Gennes energy for nematic liquid crystals in the small-elastic-constant limit. Under the uniform energy and $L^\infty$ bounds, these minimizers converge to a locally energy-minimizing harmonic map $\mathbf{Q}_0$ into a biaxial vacuum manifold. The main result of this paper is that such a limiting map $\mathbf{Q}_0$ has no interior point singularities. The proof relies on a geometric identification of the lifted Frobenius metric on the universal cover $\mathbb{S}^3$ with a rescaled Berger metric. For a hypothetical tangent cone at a point singularity, its link is a nonconstant harmonic two-sphere into the Berger sphere. We construct a smooth variation field adapted to the Hopf direction and show that it induces a quantitative instability estimate, in contradiction with the shifted stability inequality inherited from local minimality. This replaces the usual round-sphere test fields by a target-specific construction and rules out interior point defects in the biaxial setting. The result is sharp in view of the well-known existence of point defects in the uniaxial theory.

math.AP↗

Double exponential growth for the Euler equation on the whole plane

We consider the incompressible Euler equation on the whole plane $\R^2$, without boundary and without forcing, and we construct two smooth solutions whose vorticity gradient grows fast for all large times. The first solution has finite energy, integrable vorticity and initial data in every Sobolev space. Given any function $Γ(t)$ with $Γ(t)/t\to 0$, the logarithm of its vorticity gradient is not bounded by any multiple of $Γ(t)$, and, at every sufficiently large time $t$, it is larger than any multiple of $t/\log_*t$, where $\log_*$ is the iterated logarithm. The second solution has bounded vorticity, bounded vorticity derivatives on every finite time interval, and a velocity given by the absolutely convergent Biot--Savart integral which grows at most logarithmically in space. Its vorticity gradient grows double exponentially in time for all large times. Its initial vorticity is neither integrable nor square integrable, and its initial velocity and strain are unbounded. Both constructions place infinitely many localized amplification mechanisms far apart from each other in one initial datum. The first one is based on the hyperbolic stagnation point behind a smooth traveling dipole, controlled for long times by an energy--Casimir argument. The second one is based on the hyperbolic point of the Bahouri--Chemin cross on the torus, transported to the plane by spatial dilations and by the uniqueness theory of Serfati solutions.

math.AP↗

WorldFoundry: Towards Unified Infrastructure for World Intelligence

Visual generative models increasingly blur the line between synthesizing observations and simulating environments. Yet visual plausibility alone does not constitute successful world modeling: compelling outputs may lack persistent state, physical dynamics, or causally consistent responses to interactions. We define a generative world model as a predictive dynamical system that maintains a persistent, history-conditioned representation of an environment, updates that representation in response to actions or interventions, and generates temporally coherent trajectories that reflect outcomes under observed and counterfactual interventions. From this definition, we derive an intelligence-oriented framework spanning perceptual grounding, persistent state modeling, controllable transformation, dynamics-aware prediction, and closed-loop interaction. We present WorldFoundry, open-source infrastructure for reproducible inference and evaluation across video, 3D and 4D, embodied action, and interactive world systems. WorldFoundry unifies model and benchmark discovery, runtime and asset preparation, execution orchestration, artifact management, and interactive visualization, while modality-specific adapters preserve native inputs, execution semantics, output representations, and evaluation protocols. Explicit manifests, durable artifact contracts, and evidence-aware scorecards separate execution readiness from evaluative claims, making each claim traceable to its supporting artifacts. The catalog contains 270 model entries, 74 benchmarks, and 62 registered metric entries, with readiness and execution evidence recorded separately. WorldFoundry enables reproducible comparison of persistent, controllable dynamics across heterogeneous systems, advancing the field from perceptual synthesis toward measurable world intelligence.

cs.CV↗

On blow-up for the cut-off Boltzmann equation with very hard potentials

We construct smooth, strictly positive solutions of the three-dimensional cut-off Boltzmann equation with very hard potentials whose densities become unbounded in finite time. A main ingredient is a new velocity weight that is Gaussian on a ball and has a polynomial tail of fixed order. We establish coercive estimates for the linearized collision operator and bilinear estimates for the collision term, with constants independent of the radius of the Gaussian region. Combining these weighted estimates with an imploding solution of the compressible Euler equations, we develop an energy framework that controls the Boltzmann solution up to the blow-up time. The initial data agree with a constant Maxwellian outside a bounded spatial region and have finite relative entropy.

math.AP↗

Sharp global stability for the hard-sphere Boltzmann equation near equilibrium

We study the hard-sphere Boltzmann equation near a global Maxwellian and have found the bifurcation between the stability and instability in $H_{x}^{s}$ Sobolev space. We prove a strong-weak type instability result in the $s<1$ regularity regime, and establish the global stability under the regularity $s>1$ via the interplay between the dissipation and dispersive effects. In this intricate competition of dispersion, dissipation, large-velocity growth, and the asymmetric nature of the gain and loss terms, we prove that the instability is jointly generated by the dispersive effect and the loss term structure. To reach the sharp regularity for the global stability, we incorporate the dispersive effect into the dissipative energy framework for the analysis near equilibrium. This hence concludes sharp threshold and provides a complete characterization of global well-posedness near equilibrium.

math.AP↗

From entropy waves to vortex sheets for the Euler equations

We construct piecewise smooth solutions of the full compressible Euler equations that evolve from entropy waves into vortex sheets while retaining a strictly positive Taylor jump. For every integer $κ\geq1$, we obtain solutions satisfying initial material compatibility for pressure and velocity through order $κ$, whose tangential velocity jump has a nonzero leading term of order $t^{κ+1}$. These solutions are periodic in the tangential variables and admit a smooth realization in a slab with impermeable walls. We also construct solutions on a local nonperiodic tangential patch that satisfy initial material compatibility at every order, yet have a tangential velocity jump that is nonzero for every sufficiently small positive time and flat at $t=0$. The finite-order construction uses a normal Cauchy-Kowalevski problem for velocity potentials, whereas the all-order construction uses a homogeneous reduction to a smooth hyperbolic Cauchy problem.

math.AP↗

UltraDiff: Transferring High-Fidelity Priors to Compressed Latent Spaces for High-Resolution Image Generation

High-resolution synthesis with Latent Diffusion Models (LDMs) faces a fundamental trade-off between computational cost and generative fidelity. This trade off is largely governed by the latent compression factor f (e.g. f8 denotes f=8 and f32 f=32). While aggressive compression (f32) significantly improves efficiency, it suffers from severe information loss and ambiguity compared to standard f8 models. We propose UltraDiff, a framework that reformulates this challenge as joint latent diffusion learning followed by marginal score distillation. We introduce MixTeacher, a dual branch DiT that models the joint f8-f32 distribution, utilizing shared attention to resolve the ambiguity of the compressed space with high fidelity f8 priors. Crucially, we propose vector field distillation in the native f32 space. Theoretically, this compels the student to learn the conditional expectation of the teacher trajectory, effectively filtering out compression artifacts. Across the evaluated 1K to 4K settings, UltraDiff approaches f8 level quality while preserving f32 efficiency. With 20 sampling steps on a single A100 GPU, it generates a 2K image in 6.2 seconds over 40 times faster than our f8 baseline.

cs.CV↗

OmniXRef: A Unified Framework for Multimodal Multi-Reference Video Editing

Creative intent in video editing can be expressed through text instructions (e.g., "remove the people"), image references (e.g., subject appearance), video references (e.g., camera movement), or a combination of the above. We introduce OmniXRef, a unified framework for multimodal multi-reference video editing that allows users to combine variable numbers of text, image, and video references to control diverse attributes of a source video. We first devise a unified input design with a reference-bank interface and introduce $R^2$-RoPE to support generalization to more references. Then our model, OmniXRef-Model, uses a single-stream diffusion transformer to jointly process the source video and multimodal context from an MLLM in an interleaved sequence. To address the scarcity of training data for video-reference editing, we construct OmniXRef-Data using synthesis pipelines for different editing tasks and combinations of references. We use curriculum learning to progressively increase task complexity and resolution. To support systematic evaluation of video as a control modality, we introduce OmniXRef-Bench, a benchmark for video-reference attribute transfer across Camera, Motion, Content, Style, and Visual Effects (VFX). Comprehensive evaluations show that our OmniXRef-Model outperforms task-specific baselines and commercial models.

cs.CV↗

On blow-up for the supercritical defocusing nonlinear scalar wave equation

We study finite-time singularity formation for the scalar defocusing nonlinear wave equation $-\partial_t^2u+Δu=u^p$. First, for fixed $d\geq4$ and odd $p\geq3$, we establish a conditional blow-up result under the assumption that a smooth self-similar imploding profile exists for the leading-order relativistic Euler system, with associated normalized coefficients satisfying quantitative analytic radius conditions. Second, we verify both the profile and the analytic radius conditions in dimensions four and five for all sufficiently large odd $p$. Using finite speed of propagation, we obtain a dimension-independent threshold $p_0$ such that, for every $d\geq4$ and every odd $p\geq p_0$, finite-time blow-up occurs for some smooth, compactly supported, real-valued initial data. The proof combines a normalized formulation that separates the leading periodic oscillation from the unknown modulation, growth of the spatial analytic radius induced by an outgoing property, and frequency-dependent matrix multipliers that control the cumulative growth caused by the periodic linearized potential.

math.AP↗

The Inviscid Limit of Viscous Burgers through Shock Formation and Development

We study the inviscid limit of the one-dimensional viscous Burgers equation from smooth initial data through nondegenerate shock formation and subsequent shock development. Chaturvedi and Graham \cite{CG} obtained sharp estimates of order $ν^{1/4}$ up to the shock-formation time. We extend their analysis beyond this time, comparing the viscous solution with the inviscid entropy solution evolving from the same initial data. With the shock-formation time normalized to $t=0$, we prove a pointwise error bound on a fixed interval $0\leq t \leq \bar{\varepsilon}$, where $\bar{\varepsilon}>0$ is independent of the viscosity $ν$. The bound consists of a term of order $ν^{1/4}$ and an exponentially localized contribution originating from the viscous shock layer. The proof is based on matching an outer expansion, an inner expansion near the formation point, and a viscous shock-profile expansion. Combined with the results of Chaturvedi and Graham, this yields a quantitative description of the inviscid limit from the initial time through the development of the young shock.

math.AP↗

PixelDense: Dense Prediction as Representation Alignment for Pixel Diffusion

Representation alignment (REPA) accelerates diffusion transformer training, but its alignment targets are almost exclusively semantic encoders such as DINOv2 and CLIP. Recent analysis points to spatial structure, not global semantics, as the carrier of the alignment effect, yet dense-prediction foundation models trained to predict that structure remain overlooked as REPA targets. In pixel-space diffusion, SAM2, Depth Anything v2, and Metric3D v2 each outperform the DINOv2-only GenEval baseline, with the two geometric teachers leading the segmentation teacher. A flat sum of all four teachers, however, lands below the best single geometric teacher, as semantic and geometric gradients compete for one denoiser projection. We introduce PixelDense, which routes DINOv2 and SAM2 through a semantic projection stream, routes Depth Anything v2 and Metric3D v2 through a geometric projection stream, and adds a weight-space orthogonality penalty that keeps the two streams in disjoint subspaces. All four teachers are frozen during training and dropped at inference. Applied to PixelGen and DeCo with a single recipe, PixelDense improves GenEval, DPG-Bench, and HPS v2.1, raises PixelGen-XXL's GenEval Overall from 0.7927 to 0.8093, and beats every single-teacher and unfactored multi-teacher variant. In partial-noise reconstruction, independent panoptic, depth, and surface-normal probes show up to 53.1% PQ gain and 36.0% depth AbsRel reduction at $τ=0.5$ across COCO and Flickr30K. From random initialization, PixelDense also reaches the baseline's peak GenEval 1.23x faster. In SDEdit editing on PIE-Bench, PixelDense keeps more of the source background and layout at every edit strength, raising background PSNR by up to 2.2 dB.

cs.CV↗

DMA$^2$: Pixel-space Distribution Matching with Adversarial and Anchor Losses

Distribution matching distillation (DMD) provides a general framework for few-step diffusion generation, but its modern text-to-image instantiations have been developed primarily around latent diffusion. It therefore overlooks key properties and design opportunities of native RGB. We revisit two DMD interfaces for pixel-space teachers. On the teacher-matching side, diagnostics show low-noise RGB matching is dominated by a local-texture cue, motivating a fixed high-noise matching band. On the real-data side, native clean-RGB outputs allow guidance from an external visual representation without traversing a decoder or sharing the heavy fake-score critic. DINO-Adv removes this critic from the adversarial gradient path and supplies local parametric patch guidance. For distribution-level guidance, we introduce AF-Loss, a parameter-free auxiliary semantic distribution-field objective designed for text-to-image DMD. It operates on detached rolling real and generated supports in the shared DINOv2 space while preserving prompt-conditioned teacher supervision. AF-Loss adds no learnable parameters or inference-time computation. Together these designs form DMA$^2$. Across DPG-Bench, GenEval, VQAScore, and COCO30K, the four-step DMA$^2$ student performs better than the 25-step teacher and evaluated few-step distillers.

cs.CV↗

Adversarial Training for Pixel Diffusion

Pixel diffusion models generate RGB images directly, avoiding the bottleneck of an autoencoder, yet their outputs still systematically underrepresent fine-scale natural-image statistics. We show that adversarial learning provides an effective post-training correction for this deficiency. Starting from a pretrained model, we retain its original diffusion or flow-matching objective and add an adversarial loss to the predicted output at non-high-noise timesteps, leaving the model architecture and sampling procedure unchanged. To our knowledge, this is the first systematic study of adversarial post-training for pixel diffusion. Across two pixel backbones, the method jointly improves distribution fidelity, coverage, prompt alignment, and perceptual quality. We further investigate why it works. Frequency-band and power-law analyses show that the original models systematically underproduce natural-image high-frequency content, while adversarial post-training restores this missing spectral power. In contrast, perceptual loss also increases high-frequency content but sacrifices distribution fidelity and prompt alignment. Nearest-neighbor, recall, and matched no-GAN SFT controls further rule out memorization, mode dropping, and additional optimization as simple explanations. Finally, we examine the boundary of this effect. Under the tested latent diffusion configurations, the same procedure does not produce comparable joint gains and adds almost no decoded high-frequency power. These results identify direct output access to the image statistics being corrected as a key factor governing when adversarial post-training succeeds.

cs.CV↗

Sharp $H_{x}^{s}$ Ill-posedness of the Hard-sphere Boltzmann Equation

We investigate the ill-posedness mechanism of the hard-sphere Boltzmann equation in $H_{x}^{s}$ Sobolev space. Via a direct construction, we prove a strong-weak type ill-posedness result in the low-regularity regime $s<1$, establishing a sharp threshold in connection to the local $s>1$ well-posedness result [11]. Instead of originating from the large-velocity growth of the collision kernel, this illposedness is generated by the loss term and dispersive effects. Consequently, we prove a dispersion-driven nonlinear instability mechanism for the hard-sphere Boltzmann equation, and provide a capstone of the ill-posedness series [18,20].

math.AP↗

SupGRPO: Enhancing GRPO with Matching-based Online SFT for Text Spotting

Text spotting requires both accurate text recognition and precise spatial localization. Current specialised spotters excel at predicting tight bounding boxes in natural scenes, but falter on complex or artistic text, whereas multimodal large language models (MLLMs) possess strong recognition capabilities yet remain weak at localisation. To equip the text spotter with general and powerful recognition capabilities and to maximize its localization ability, we explore two MLLM-based fine-tuning methods: Supervised Fine-Tuning (SFT) and reinforcement learning fine-tuning based on Group Relative Policy Optimisation (GRPO). An interesting finding is that SFT is less effective than GRPO at enhancing recognition, while GRPO is less effective than SFT at enhancing detection. To compensate for each other's shortcomings, we introduce a joint training strategy, SupGRPO, which simultaneously optimizes the model using both SFT and GRPO. SupGRPO employs the specially designed reward functions and develops a matching-based online SFT applied solely to coordinate tokens. It both mitigates the reward sparsity problem of GRPO and avoids the instance order dependency problem of SFT. To evaluate particularly challenging cases, we curate ATS, a dataset for artistic text spotting. Experiments demonstrate that SupGRPO improves both text recognition and detection, and attains superior performance. Our code and dataset will be released at https://github.com/Psycho-9/SupGRPO.

cs.CV↗

Exponential mixing for the stochastic Navier--Stokes equation with localized noise

This paper studies the 2D Navier--Stokes equation on a bounded domain with Navier-slip boundary conditions, driven by spatially localized stationary forcing generated by a stochastic heat equation. We prove exponential mixing for the associated Navier--Stokes--heat system, and consequently exponential convergence of the velocity law under stationary forcing. The main difficulty is that the white noise is confined to a subdomain and reaches the velocity indirectly through the heat component. To address the degeneracy, the proof combines Malliavin calculus with PDE control theory. The key ingredient is a stabilization scheme that converts localized Navier--Stokes controls into time-regular controls compatible with the heat dynamics.

math.PR↗

Benchmarking Clinical Decision Pathway Adherence in Large Language Models

Following clinical decision pathways (CDPs) defined by clinical practice guidelines is essential for safe and reliable medical decision-making. However, existing medical large language model (LLM) benchmarks mainly evaluate final-answer accuracy, providing limited evaluation of models' ability to adhere to guidelines. To address this gap, we introduce MEGA-CDP, a benchmark for evaluating whether medical LLMs can generate guideline-adherent CDPs using provided guidelines as references. MEGA-CDP is constructed from 2,274 English and Chinese clinical practice guidelines through a guideline-to-case pipeline, yielding 42,353 clinical cases with explicit reference CDPs. It supports both single-turn vignette and multi-turn interactive settings, and introduces a CDP-oriented evaluation framework for measuring pathway consistency. Experiments on 16 representative LLMs show that reliable clinical decision support remains challenging for current models, demonstrating the need for CDP-oriented evaluation and the value of MEGA-CDP for advancing guideline adherence in medical LLMs.

cs.CL↗