arXiv subjects

Junhao Zhang

Publications and source records attributed to Junhao Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Fair on the Surface? Benchmarking Hidden-Output Fairness Gaps in LLM Recommenders

Fairness audits for LLM-based recommenders have largely focused on observable outputs, implicitly assuming that stable recommendations reflect stable internal processing. We challenge this assumption with FairGap, the first benchmark to jointly evaluate recommendation fairness at two levels: observable output shift (OBS) and hidden representation shift (IBS), measured through controlled counterfactual identity probes across gender, age, and race. Their relationship is summarized via Representation-Output Alignment (ROA), with quadrant diagnostics for identifying user-level hidden-output mismatch. Applied to six open-weight LLM families across three domains, FairGap reveals pervasive hidden-output decoupling: ROA rarely exceeds 0.22, and a non-negligible user population shows stable outputs despite substantial internal shifts, a mode that output-only audits cannot detect by design. Further, activation steering that reduces IBS by up to 8x simultaneously worsens OBS, demonstrating a fundamental tension between internal and output-level fairness that existing frameworks are unequipped to diagnose.

cs.AI↗

arXiv2026

ScaleResfusion: Residual Rectified Flow based on Residual Vector Field

Real-world Image Restoration (Real-IR) aims to recover high-quality (HQ) images from complex and unknown degradations. Recent diffusion-based methods have substantially improved perceptual quality, yet two obstacles remain: methods that sample from Gaussian noise require many steps and are often less faithful to the degraded input, whereas residual-based methods that start from the low-quality (LQ) image typically train task-specific models from scratch, with optimization objectives coupled to a particular noise scheduler, and therefore cannot reuse modern pre-trained generative priors. We present \textbf{ScaleResfusion}, which rewrites residual restoration as a scheduler-independent adaptation interface for pre-trained text-to-image rectified-flow models. Its core, \textbf{Residual Rectified Flow} (RRF), inserts the residual term $R$ into the linear transport path of Rectified Flow, so that sampling starts from noisy LQ at an exact acceleration point, where the signal-to-noise ratio of the starting state is continuously controlled by the residual ratio $\gamma$. The resulting optimization target, the \textbf{residual vector field}, contains no scheduler-specific coefficients and differs from the pre-trained rectified-flow target only by the residual offset $\gamma R$; adapting a frozen billion-scale backbone therefore reduces to fitting this compact residual correction with LoRA-only training. A knowledge-distillation pipeline built around RRF further reduces sampling to as few as 4 steps. Experiments on real-world super-resolution across multiple benchmarks show that ScaleResfusion achieves state-of-the-art restoration quality and transfers consistently across pre-trained rectified-flow backbones from 2B to 9B parameters.

cs.CV↗

arXiv2026

The SPOTLIGHT Pulsar Search Pipeline: A GPU-Accelerated FFT Approach

We present the pulsar search component of SPOTLIGHT (Survey for sPoradic radiO bursTs via a commensaL multI-beam Gpu-powered Hpc at the gmrT), a GPU-accelerated commensal backend operating at the upgraded Giant Metrewave Radio Telescope (uGMRT). While SPOTLIGHT is primarily designed for real-time detection and localisation of fast radio bursts (FRBs), it simultaneously records a subset of beamformed data products for periodicity searches without requiring dedicated telescope time. To process the large data volumes generated by the survey, we have developed a scalable FFT-based pulsar search pipeline that combines radio-frequency interference mitigation, GPU-accelerated dedispersion and periodicity searches, multi-beam candidate sifting, efficient folding and machine-learning classification. Using population synthesis and archival uGMRT observations, we estimate that a fully operational SPOTLIGHT survey with 160 PC and one IA beam could discover $\sim$ 450 new pulsars, probing both high-sky coverage and faint pulsars over three and a half years of commensal observations. The pipeline has been validated on GMRT Cycle 48 and 49 observations (i.e. April 2025 to Mar 2026), successfully re-detecting numerous known pulsars with a wide range of Period, DM and flux densities, and is currently operational for SPOTLIGHT commensal data processing. We describe the SPOTLIGHT observing system, pulsar survey design, search parameter space, candidate-selection strategy, current status, and future developments. SPOTLIGHT demonstrates the scientific potential of commensal pulsar surveys and serves as a pathfinder for real-time, large-scale pulsar and transient searches in the SKA era.

astro-ph.HE↗

arXiv2026

$L^{2}-L^{\infty}$ decay estimates and inviscid limits for Global smooth solutions to the compressible Navier-Stokes-Riesz system

We study the Cauchy problem in $\mathbb{R}^{3}$ for the repulsive compressible Navier-Stokes-Riesz system with Riesz exponent $0<s<1$ and viscosity $0<\varepsilon\leq1$, where the Riesz interaction $\nabla(-\Delta)^{-s}(\rho-\bar{\rho})$ is a generalization of the Coulomb interaction for electrons. For small perturbations of a constant equilibrium, with the solenoidal component of the initial velocity of order $\mathcal{O}(\varepsilon)$, we prove the global existence and uniqueness of smooth solutions. We derive time-decay estimates in $L^{2}$ norms and $L^{\infty}$ norms that capture both uniform-in-$\varepsilon$ dispersive behavior and viscosity-dependent dissipation. We further establish a global-in-time inviscid limit to the irrotational global solution of the compressible Euler-Riesz system whose initial data consist of the same density and the curl-free component of the velocity, with an explicit convergence rate in $W^{k,p}$ norms. The proof combines viscosity-adapted dispersive estimates, normal-form analysis and nonlinear energy estimates with control of both negative and positive Sobolev norms.

math.AP↗

arXiv2026

Runtime-Orchestrated Second-Order Optimization for Scalable LLM Training

Second-order methods offer an attractive path toward more sample-efficient LLM training, but their practical use is often blocked by the systems cost of maintaining and updating large matrix-based optimizer states. We introduce \textbf{Asteria}, a runtime system designed to remove this bottleneck by separating second-order optimization logic from the critical GPU training path. Rather than keeping all preconditioner state on the accelerator, Asteria dynamically distributes optimizer state across GPU memory, CPU memory, and optional NVMe storage according to architectural constraints and runtime pressure. It further uses training hooks to prepare shadow states in advance, allowing expensive inverse-root computations to proceed asynchronously on the host while GPU computation continues. For distributed training, Asteria employs a bounded-staleness protocol that limits synchronization frequency while preserving optimizer effectiveness through topology-aware coordination. We evaluate Asteria on both memory-constrained and distributed training settings. On a DGX Spark platform with a single GB10 GPU and 128GB unified memory, Asteria supports second-order training for a 1B-parameter language model. On multi-node GH200 systems, it lowers visible optimizer overhead, reduces recurring latency spikes, accelerates convergence in wall-clock time, and maintains the optimization advantages of SOAP and KL-Shampoo in a 7B-parameter language model. Our results suggest that second-order LLM training can be made practical not by simplifying the optimizer alone, but by rethinking how optimizer state, background computation, and distributed synchronization are managed at the runtime level.

cs.DC↗

arXiv2026

UCPO: Uncertainty-Aware Policy Optimization

The key to building trustworthy large language models (LLMs) lies in endowing them with inherent uncertainty expression capabilities, thereby mitigating overconfident errors in high-stakes applications. However, existing RL paradigms such as GRPO often suffer from Advantage Bias due to binary decision spaces and static uncertainty rewards, inducing either excessive conservatism or overconfidence. To tackle this challenge, this paper unveils the root causes of reward hacking and overconfidence in current RL paradigms incorporating uncertainty-based rewards, based on which we propose the UnCertainty-Aware Policy Optimization (UCPO) framework. UCPO employs Ternary Advantage Decoupling to separate and independently normalize deterministic and uncertain rollouts, thereby eliminating advantage bias. Furthermore, a Dynamic Uncertainty Reward Adjustment mechanism adapts uncertainty weights in real-time according to model evolution and instance difficulty. Experimental results in mathematical reasoning and general tasks demonstrate that UCPO effectively resolves the reward imbalance, significantly improving the reliability of the model beyond their knowledge boundaries.

cs.AI↗

arXiv2026

Asymptotic stability of steady states for the compressible Navier-Stokes-Riesz system in the presence of vacuum

We consider a one-dimensional physical vacuum free boundary problem on the compressible Navier-Stokes-Riesz system for an attractive Riesz potential $|x|^{2s-1}/(2s-1)$ with $0<s<1/2$. It is proved that for the adiabatic constant $\gamma$ satisfying $2(1-s)<\gamma<1+2s/3$ under the additional condition that $3/8<s<1/2$, there exists a unique global-in-time strong solution. Specifically, we establish the Lyapunov-type stability of the compactly supported steady states in the Lagrangian coordinates and we also obtain the time rate of convergence for the strong solution to steady states with the same mass in weighted Sobolev spaces where the weights indicate the behavior of solutions near the vacuum free boundary. The difficulties and challenges in the proof are caused not only by the degeneracy due to the vacuum free boundary but also by the non-local feature of the Riesz potential.

math.AP↗

arXiv2025

Atmospheric Turbulence-Resilient Long-Range Fourier Ptychography

While Fourier ptychography (FP) offers super-resolution for macroscopic imaging, its real-world application is severely hampered by atmospheric turbulence, a challenge largely unaddressed in existing macroscopic FP research operating under idealized conditions. This work establishes, to our knowledge, the first comprehensive computational framework specifically designed for turbulence mitigation in long-range FP, termed Turbulence-Mitigated FP (TMFP). Rather than correcting pupil errors, an image degradation model is developed alongside a reconstruction pipeline inspired by speckle interferometry. By taking multiple short-exposure randomly-distorted measurements and exploiting their statistical properties, the diffraction-limited sub-aperture images can be recovered for further FP reconstruction. Numerical simulations and experimental validations under optical turbulence demonstrate the method's robustness, resolution enhancement, and practicality in adverse conditions, paving the way for the reliable deployment of high-resolution macroscopic FP in real-world scenarios.

physics.optics↗

arXiv2025

Hybrid RIS-Enhanced ISAC Secure Systems: Joint Optimization in the Presence of an Extended Target

Unlike the conventional fully-passive and fully-active reconfigurable intelligent surfaces (RISs), a hybrid RIS consisting of active and passive reflection units has recently been concerned, which can exploit their integrated advantages to alleviate the RIS-induced path loss. In this paper, we investigate a novel security strategy where the multiple hybrid RIS-aided integrated sensing and communication (ISAC) system communicates with downlink users and senses an extended target synchronously. Assuming imperfectly known target location (TL), we consider the joint design of the transmit signal and receive filter bank of the base station (BS), the receive beamformers of all users and the weights of the hybrid RIS. An optimization problem is formulated for maximizing the worst-case sensing signal-to-interference-plus-noise-ratio (SINR) subject to secure communication and system power budget constraints. To address this non-convex problem, we leverage generalized fractional programming (GFP) and penalty-dual-decomposition (PDD), and propose a security solution that efficiently optimizes all variables by employing convex optimization approaches. Simulation results show that by incorporating the multiple hybrid RIS into the optimization design, the extended target detection and secure transmission performance of ISAC systems are improved over the state-of-the-art RIS-aided ISAC approaches.

physics.ins-det↗

arXiv2025

Lost-in-the-Middle in Long-Text Generation: Synthetic Dataset, Evaluation Framework, and Mitigation

Existing long-text generation methods primarily concentrate on producing lengthy texts from short inputs, neglecting the long-input and long-output tasks. Such tasks have numerous practical applications while lacking available benchmarks. Moreover, as the input grows in length, existing methods inevitably encounter the "lost-in-the-middle" phenomenon. In this paper, we first introduce a Long Input and Output Benchmark (LongInOutBench), including a synthetic dataset and a comprehensive evaluation framework, addressing the challenge of the missing benchmark. We then develop the Retrieval-Augmented Long-Text Writer (RAL-Writer), which retrieves and restates important yet overlooked content, mitigating the "lost-in-the-middle" issue by constructing explicit prompts. We finally employ the proposed LongInOutBench to evaluate our RAL-Writer against comparable baselines, and the results demonstrate the effectiveness of our approach. Our code has been released at https://github.com/OnlyAR/RAL-Writer.

cs.CL↗

arXiv2025

Ground State of $\mathrm{SU}\left(3\right)$ spin model on the checkerboard lattice

Geometric frustration in quantum spin systems can lead to exotic ground states. In this study, we investigate the $\mathrm{SU}(3)$ spin model on the checkerboard lattice to explore the effects of frustration arising from its point-connected $(N+1)$-site local structure. We employ density matrix renormalization group (DMRG) and exact diagonalization (ED) techniques to determine the ground state properties. Our results reveal the absence of both 3-sublattice antiferromagnetic order and valence cluster solid order. Instead, we identify ground states with bond stripe patterns sensitive to boundary conditions and system size, comprising staggered singlet arrays and uniform flat stripes. Notably, these stripes are relatively decoupled, and similar patterns can be reconstructed in quasi-one-dimensional ladders. These findings suggest that geometric frustration drives the system toward a mixed phase, combining characteristics of spin-liquid and valence cluster solid states, providing new insights into the behavior of frustrated quantum spin systems.

cond-mat.str-el↗

arXiv2024

Automatic State Machine Inference for Binary Protocol Reverse Engineering

Protocol Reverse Engineering (PRE) is used to analyze protocols by inferring their structure and behavior. However, current PRE methods mainly focus on field identification within a single protocol and neglect Protocol State Machine (PSM) analysis in mixed protocol environments. This results in insufficient analysis of protocols' abnormal behavior and potential vulnerabilities, which are crucial for detecting and defending against new attack patterns. To address these challenges, we propose an automatic PSM inference framework for unknown protocols, including a fuzzy membership-based auto-converging DBSCAN algorithm for protocol format clustering, followed by a session clustering algorithm based on Needleman-Wunsch and K-Medoids algorithms to classify sessions by protocol type. Finally, we refine a probabilistic PSM algorithm to infer protocol states and the transition conditions between these states. Experimental results show that, compared with existing PRE techniques, our method can infer PSMs while enabling more precise classification of protocols.

cs.CR↗

arXiv2024

Global dynamics of isothermal rarefied gas flows in an infinite layer

Let rarefied gas be confined in an infinite layer with diffusely reflecting boundaries that are isothermal and non-moving. The initial-boundary value problem on the nonlinear Boltzmann equation governing the rarefied gas flow in such setting is challenging due to unboundedness of both domain and its boundaries as well as the presence of physical boundary conditions. In the paper, we establish the global-in-time dynamics of such rarefied gas flows near global Maxwellians in three or two-dimensions. For the former case, we also prove that the solutions decay in time at a polynomial rate which is the same as that of solutions to the two-dimensional heat equation. This is the first result on global solutions of the Boltzmann equation with non-compact and diffuse boundaries.

math.AP↗

arXiv2024

EvolveDirector: Approaching Advanced Text-to-Image Generation with Large Vision-Language Models

Recent advancements in generation models have showcased remarkable capabilities in generating fantastic content. However, most of them are trained on proprietary high-quality data, and some models withhold their parameters and only provide accessible application programming interfaces (APIs), limiting their benefits for downstream tasks. To explore the feasibility of training a text-to-image generation model comparable to advanced models using publicly available resources, we introduce EvolveDirector. This framework interacts with advanced models through their public APIs to obtain text-image data pairs to train a base model. Our experiments with extensive data indicate that the model trained on generated data of the advanced model can approximate its generation capability. However, it requires large-scale samples of 10 million or more. This incurs significant expenses in time, computational resources, and especially the costs associated with calling fee-based APIs. To address this problem, we leverage pre-trained large vision-language models (VLMs) to guide the evolution of the base model. VLM continuously evaluates the base model during training and dynamically updates and refines the training dataset by the discrimination, expansion, deletion, and mutation operations. Experimental results show that this paradigm significantly reduces the required data volume. Furthermore, when approaching multiple advanced models, EvolveDirector can select the best samples generated by them to learn powerful and balanced abilities. The final trained model Edgen is demonstrated to outperform these advanced models. The code and model weights are available at https://github.com/showlab/EvolveDirector.

cs.CV↗

arXiv2024

Steady compressible Navier-Stokes-Fourier system with slip boundary conditions arising from kinetic theory

This paper studies the boundary value problem on the steady compressible Navier-Stokes-Fourier system in a channel domain $(0,1)\times\mathbb{T}^2$ with a class of generalized slip boundary conditions that were systematically derived from the Boltzmann equation by Coron \cite{Coron-JSP-1989} and later by Aoki et al \cite{Aoki-Baranger-Hattori-Kosuge-Martalo-Mathiaud-Mieussens-JSP-2017}. We establish the existence and uniqueness of strong solutions in $(L_{0}^{2}\cap H^{2}(\Omega))\times V^{3}(\Omega)\times H^{3}(\Omega)$ provided that the wall temperature is near a positive constant. The proof relies on the construction of a new variational formulation for the corresponding linearized problem and employs a fixed point argument. The main difficulty arises from the interplay of velocity and temperature derivatives together with the effect of density dependence on the boundary.

math.AP↗

arXiv2024

Ground State Phase Diagram of $\text{SU}(3)$ $t$-$J$ Chain

Distinct from the $\text{SU}(2)$ case, the fermionic systems with $\text{SU}(N)$ symmetry are expected to exhibit novel physics, such as exotic singlet formation. Using the density matrix renormalization group technique, we obtain the ground state phase diagram of the $\text{SU}(3)$ $t$-$J$ chain for density $n<1$. The ground state phase diagram includes the Luttinger liquid, the extended Luther-Emery liquid characterized by a spin gap, and the phase separation state. We quantitatively assess the characteristics of the three phases by measuring spin gap, compressibility, various correlation functions and structure factors. We further study the extended Luther-Emery liquid phase and discover molecular superfluid quasi-long-range order. The mechanism of the molecular superfluid is the combination of three $\text{SU}(3)$ fermions on sites that are not completely connected. Accordingly, we can speculate the behavior of the $\text{SU}(N)$ $t$-$J$ chain model with larger $N$ values, operating within the same filling regime.

cond-mat.str-el↗

arXiv2024

ConStyle v2: A Strong Prompter for All-in-One Image Restoration

This paper introduces ConStyle v2, a strong plug-and-play prompter designed to output clean visual prompts and assist U-Net Image Restoration models in handling multiple degradations. The joint training process of IRConStyle, an Image Restoration framework consisting of ConStyle and a general restoration network, is divided into two stages: first, pre-training ConStyle alone, and then freezing its weights to guide the training of the general restoration network. Three improvements are proposed in the pre-training stage to train ConStyle: unsupervised pre-training, adding a pretext task (i.e. classification), and adopting knowledge distillation. Without bells and whistles, we can get ConStyle v2, a strong prompter for all-in-one Image Restoration, in less than two GPU days and doesn't require any fine-tuning. Extensive experiments on Restormer (transformer-based), NAFNet (CNN-based), MAXIM-1S (MLP-based), and a vanilla CNN network demonstrate that ConStyle v2 can enhance any U-Net style Image Restoration models to all-in-one Image Restoration models. Furthermore, models guided by the well-trained ConStyle v2 exhibit superior performance in some specific degradation compared to ConStyle.

cs.CV↗

arXiv2024

ACTION: Augmentation and Computation Toolbox for Brain Network Analysis with Functional MRI

Functional magnetic resonance imaging (fMRI) has been increasingly employed to investigate functional brain activity. Many fMRI-related software/toolboxes have been developed, providing specialized algorithms for fMRI analysis. However, existing toolboxes seldom consider fMRI data augmentation, which is quite useful, especially in studies with limited or imbalanced data. Moreover, current studies usually focus on analyzing fMRI using conventional machine learning models that rely on human-engineered fMRI features, without investigating deep learning models that can automatically learn data-driven fMRI representations. In this work, we develop an open-source toolbox, called Augmentation and Computation Toolbox for braIn netwOrk aNalysis (ACTION), offering comprehensive functions to streamline fMRI analysis. The ACTION is a Python-based and cross-platform toolbox with graphical user-friendly interfaces. It enables automatic fMRI augmentation, covering blood-oxygen-level-dependent (BOLD) signal augmentation and brain network augmentation. Many popular methods for brain network construction and network feature extraction are included. In particular, it supports constructing deep learning models, which leverage large-scale auxiliary unlabeled data (3,800+ resting-state fMRI scans) for model pretraining to enhance model performance for downstream tasks. To facilitate multi-site fMRI studies, it is also equipped with several popular federated learning strategies. Furthermore, it enables users to design and test custom algorithms through scripting, greatly improving its utility and extensibility. We demonstrate the effectiveness and user-friendliness of ACTION on real fMRI data and present the experimental results. The software, along with its source code and manual, can be accessed online.

eess.IV↗