arXiv subjects

Rui Zhang

Publications and source records attributed to Rui Zhang.

At least 163 records · Page 9Linked to original sources

arXiv2026

Slow Movable Antenna System Design Based on Cell-Specific Long-Term Angular Power Spectrum

Movable antenna (MA) has recently emerged as a promising paradigm for enhancing wireless communication performance by exploiting spatial degrees of freedom through flexible antenna repositioning. However, most existing designs rely on short-term user-specific instantaneous/statistical channel state information (CSI), which incurs excessive channel estimation overhead and complexity due to frequent antenna movement. To address this issue, this paper proposes a new design framework for antenna position optimization over a much longer timescale based on the cell-level statistical channel information acquired at the base station (BS). To this end, a cell-specific statistical channel model is developed for MA-aided multiuser communication systems, based on which the antenna position optimization framework for maximizing the ergodic system utility is formulated. Then, the covariance-eigenvalues-balancing antenna positions (CEBAP) design is derived to asymptotically approximate optimal solutions by statistically reducing users' channel correlation. Notably, the CEBAP solution solely depends on the BS-side angular power spectrum (APS) of wireless channels for mobile users across the cell, which significantly alleviates the overhead of channel acquisition and antenna movement, and yet remains effective for improving various system utilities over long timescales, such as weighted sum rate and minimum signal-to-interference-plus-noise ratio. Moreover, a low-complexity log-barrier penalized optimization (LOBPO) method is proposed to numerically solve the CEBAP. Simulation results based on realistic urban layouts and ray-tracing channels demonstrate consistent performance gains of the proposed CEBAP over fixed-position antenna systems across different utility functions, which closely approaches the upper bound achieved by instantaneous CSI-based MA optimization for moderately large antenna regions.

eess.SP↗

arXiv2026

When Relations Break: Analyzing Relation Hallucination in Vision-Language Model Under Rotation and Noise

Vision-language models (VLMs) achieve strong multimodal performance but remain prone to relation hallucination, which requires accurate reasoning over inter-object interactions. We study the impact of visual perturbations, specifically rotation and noise, and show that even mild distortions significantly degrade relational reasoning across models and datasets. We further evaluate prompt-based augmentation and preprocessing strategies (orientation correction and denoising), finding that while they offer partial improvements, they do not fully resolve hallucinations. Our results reveal a gap between perceptual robustness and relational understanding, highlighting the need for more robust, geometry-aware VLMs.

cs.CV↗

arXiv2026

Holographic Surface Enabled Integrated Sensing and Communications

Integrated sensing and communications (ISAC) is an essential 6G capability for joint data transmission and environmental sensing. To support 6G scenarios with stringent ISAC performance requirements, existing massive-MIMO-based systems are expected to scale toward ultra-massive MIMO. However, this scaling incurs prohibitive cost and power consumption when realized using widely adopted phased arrays with complex phase shifters and feeding networks. Recently, holographic integrated sensing and communications (HISAC) has emerged as a promising paradigm to address this issue. It employs reconfigurable holographic surfaces (RHSs), a type of leaky-wave antenna, as a cost- and energy-efficient implementation of ultra-massive MIMO-based ISAC, and offers enhanced flexibility for ISAC beam synthesis through holographic beamforming. In this paper, we provide a comprehensive tutorial on HISAC, focusing on how RHS-enabled holographic beamforming can be exploited to jointly support communication and sensing under practical hardware constraints. We first introduce the fundamentals of RHSs and discuss the unique leakage power constraint of holographic beamforming. We then present a general optimization framework for HISAC and show how HISAC enhances joint communication and sensing, sensing-assisted communication, and communication-assisted sensing. We further present HISAC system implementations and experimental results. Finally, we outline promising research directions for HISAC, highlighting the potential of HISAC in advancing efficient, flexible, and high-performance ISAC networks.

eess.SP↗

arXiv2026

SilverTorch: A Unified Model-based System to Democratize Large-Scale Recommendation on GPUs

Serving deep learning based recommendation models (DLRM) at scale is challenging. Existing approaches rely on dedicated ANN indexing and filtering services on CPUs, suffering from non-negligible costs and missing co-design opportunities. Such inefficiency makes them difficult to support complex model architectures, such as learned similarities and multi-task retrieval. In this paper, we present SilverTorch, a model-based serving system that brings all components into one unified model. It unifies model serving by replacing standalone indexing and filtering services with model layers. We propose a model-based GPU Bloom index for feature filtering and a fused Int8 ANN kernel for nearest neighbor search. Through co-design of the ANN search and feature filtering, we reduce GPU memory usage and eliminate computation. Benefiting from this design, we scale up retrieval by introducing an OverArch scoring layer and a multi-task retrieval with a Value Model to aggregate scores. These advancements improve the retrieval accuracy and enable future studies for serving more complex models. Our evaluation on industry-scale datasets show that SilverTorch achieves up to 23.7\times higher throughput compared to the state-of-the-art approaches. We also demonstrate that SilverTorch solution is 13.35\times more cost-efficient than CPU-based solution while improving accuracy via serving more complex models. SilverTorch is deployed at scale, serving hundreds of models online and supporting recommendation for diverse applications.

cs.IR↗

arXiv2026

eLLM: Elastic Memory Management Framework for Efficient LLM Serving

Large Language Models are increasingly being deployed in datacenters. Serving these models requires careful memory management, as their memory usage includes static weights, dynamic activations, and key-value caches. While static weights are constant and predictable, dynamic components such as activations and KV caches change frequently during runtime, presenting significant challenges for efficient memory management. Modern LLM serving systems typically handle runtime memory and KV caches at distinct abstraction levels: runtime memory management relies on static tensor abstractions, whereas KV caches utilize a page table-based virtualization layer built on top of the tensor abstraction. This virtualization dynamically manages KV caches to mitigate memory fragmentation. However, this dual-level approach fundamentally isolates runtime memory and KV cache management, resulting in suboptimal memory utilization under dynamic workloads, which can lead to a nearly 20% drop in throughput. To address these limitations, we propose eLLM, an elastic memory management framework inspired by the classical memory ballooning mechanism in operating systems. The core components of eLLM include: (1) Virtual Tensor Abstraction, which decouples the virtual address space of tensors from the physical GPU memory, creating a unified and flexible memory pool; (2) an Elastic Memory Mechanism that dynamically adjusts memory allocation through runtime memory inflation and deflation, leveraging CPU memory as an extensible buffer; and (3) a Lightweight Scheduling Strategy employing SLO-aware policies to optimize memory utilization and effectively balance performance trade-offs under stringent SLO constraints. Comprehensive evaluations demonstrate that eLLM significantly outperforms state-of-the-art systems, 2.32x higher decoding throughput, and supporting 3x larger batch sizes for 128K-token inputs.

cs.DC↗

arXiv2026

QHap: Quantum-Inspired Haplotype Phasing

Haplotype phasing, the process of resolving parental allele inheritance patterns in diploid genomes, is critical for precision medicine and population genetics, yet the underlying optimization is NP-hard, posing a scalability challenge. To address this, we introduce QHap, a haplotype phasing algorithm that leverages quantum-annealing-inspired optimization. By reformulating haplotype phasing as a Max-Cut problem and deploying a GPU-accelerated ballistic simulated bifurcation solver, QHap accelerates phasing while maintaining accuracy comparable to established phasing tools. On the highly polymorphic human major histocompatibility complex region, QHap demonstrates 4- to 20-fold acceleration over HapCUT2 and WhatsHap with zero switch error across multiple long-read sequencing platforms. The framework implements two strategies: a read-based method for regional phasing, and a single nucleotide polymorphism-based method that, through quality-weighted probabilistic edge construction, efficiently scales to chromosome-scale tasks. Integration of Pore-C chromatin conformation capture data increases the haplotype N50 by up to 15-fold, enabling near-chromosome-scale haplotype reconstruction. QHap demonstrates that quantum-inspired algorithms operating on classical hardware offer a promising approach to addressing the growing computational demands of sequencing data, establishing a new paradigm for applying physics-inspired optimization to fundamental challenges in computational genomics.

q-bio.GN↗

arXiv2026

Joint Communication and Trajectory Design for Movable Antenna Systems

Movable antennas (MAs) have attracted significant attention in wireless communications due to their ability to reconfigure channel conditions by flexibly adjusting the antenna positions within a confined region. However, MA movement generally incurs a non-negligible delay, which may significantly limit the data transmission time at optimized positions. To tackle this challenge, this paper investigates a new joint communication and trajectory optimization problem, where each MA transmits while moving along an optimized trajectory to prolong the effective data transmission time. Focusing on a single-MA system, our goal is to maximize the average data rate by optimizing the MA's positions over time, subject to its maximum velocity constraints. However, this continuous-time antenna position optimization problem is highly non-convex and challenging to solve. To tackle this challenge, we first consider a special case with two channel paths and derive the optimal MA trajectory in closed form. For other general cases, we ingeniously reformulate the average rate maximization problem into a fixed-hop shortest path problem in graph theory by sampling the antenna movement region into a multitude of discrete points, and solve it optimally. Simulation results demonstrate that our proposed algorithm can significantly improve the data rate compared to other baseline schemes.

eess.SP↗

arXiv2026

Parton Distribution Functions from Large Momentum Expansion of Current-Current Correlators

The universality of the large momentum expansion allows computing parton distribution functions (PDFs) starting from any Euclidean correlator with appropriate large momentum Fourier Components. Here we consider current-current correlators which have been used in short-distance expansion to obtain moments of PDFs. The advantage of such correlators is that they have simple renormalization properties and do not have linear power divergences as in quasi-PDF. However, in lattice calculations, four-point functions are needed. Here we present an expansion formula with current-current correlators up to the next-to-leading order, and preliminary numerical calculations with four-point functions.

hep-ph↗

arXiv2026

SAMamba3D: adapting Segment Anything for generalizable 3D segmentation of multiphase pore-scale images

Reliable segmentation of multiphase pore-scale X-ray images of rocks is necessary to quantify fluid saturation, connectivity, and interfacial geometry. However, current 3D segmentation methods are typically dataset-specific, requiring retraining or extensive fine-tuning whenever rock type, fluid pattern, scanner, or acquisition conditions change. Foundation models such as the Segment Anything Model (SAM) provide strong 2D boundary priors, but they are not directly applicable to 3D data. We present SAMamba3D, a parameter-efficient framework that adapts a largely frozen SAM encoder to generalizable 3D pore-scale segmentation by coupling it with Mamba-based volumetric context modeling and progressive cross-scale feature interaction. For sandstone and carbonate datasets, with different fluids, wettability, and scanning conditions, SAMamba3D matches or outperforms current 3D baselines while reducing the need for case-specific retraining. The resulting segmented images preserve physically meaningful descriptors, including fluid saturation, connectivity, and interface morphology, enabling more reliable and rapid analysis of large 3D multiphase images.

cs.CV↗

arXiv2026

AutoPPA: Automated Circuit PPA Optimization via Contrastive Code-based Rule Library Learning

Performance, power, and area (PPA) optimization is a fundamental task in RTL design, requiring a precise understanding of circuit functionality and the relationship between circuit structures and PPA metrics. Recent studies attempt to automate this process using LLMs, but neither feedback-based nor knowledge-based methods are efficient enough, as they either design without any prior knowledge or rely heavily on human-summarized optimization rules. In this paper, we propose AutoPPA, a fully automated PPA optimization framework. The key idea is to automatically generate optimization rules that enhance the search for optimal solutions. To do this, AutoPPA employs an Explore-Evaluate-Induce ($E^2I$) workflow that contrasts and abstracts rules from diverse generated code pairs rather than manually defined prior knowledge, yielding better optimization patterns. To make the abstracted rules more generalizable, AutoPPA employs an adaptive multi-step search framework that adopts the most effective rules for a given circuit. Experiments show that AutoPPA outperforms both the manual optimization and the state-of-the-art methods SymRTLO and RTLRewriter.

cs.LG↗

arXiv2026

Wireless Communication for Low-Altitude Economy with UAV Swarm Enabled Two-Level Movable Antenna System

Unmanned aerial vehicle (UAV) is regarded as a key enabling platform for low-altitude economy, due to its advantages such as 3D maneuverability, flexible deployment, and LoS air-to-air/ground communication links. In particular, the intrinsic high mobility renders UAV especially suitable for operating as a movable antenna (MA) from the sky. In this paper, by exploiting the flexible mobility of UAV swarm and antenna position adjustment of MA, we propose a novel UAV swarm enabled two-level MA system, where UAVs not only individually deploy a local MA array, but also form a larger-scale MA system with their individual MA arrays via swarm coordination. We formulate a general optimization problem to maximize the minimum achievable rate over all ground user equipments (UEs), by jointly optimizing the 3D UAV swarm placement positions, their individual MAs' positions, and receive beamforming for different UEs. To gain useful insights, we first consider the special case where each UAV has only one antenna, under different scenarios of one single UE, two UEs, and arbitrary number of UEs. In particular, for the two-UE case, we derive the optimal UAV swarm placement positions in closed-form that achieves IUI-free communication when the uniform plane wave (UPW) model holds, where the UAV swarm forms a uniform sparse array (USA) satisfying minimum safe distance constraint. While for the general case with arbitrary number of UEs, we propose an efficient alternating optimization algorithm to solve the formulated non-convex optimization problem. Then, we extend the results to the case where each UAV is equipped with multiple antennas. Numerical results verify that the proposed low-altitude UAV swarm enabled MA system significantly outperforms various benchmark schemes, thanks to the exploitation of two-level mobility to create more favorable channel conditions for multi-UE communications.

cs.IT↗

arXiv2026

Black-Box Skill Stealing Attack from Proprietary LLM Agents: An Empirical Study

Large language model (LLM) agents increasingly rely on skills to package reusable capabilities through instructions, tools, and resources. High-quality skills embed expert knowledge, curated workflows, and execution constraints into agents, fueling a growing skill economy through their value and scalability. Yet this ecosystem also creates a new attack surface, as adversaries can interact with public agent interfaces to extract hidden proprietary skill content. We present the first systematic study of black-box skill stealing against LLM agent systems. Compared with conventional system prompt stealing, skill stealing targets modular and structured capability packages whose leakage is directly actionable for copying, redistribution, and monetization, making the resulting harm potentially greater. To study this threat, we derive an attack taxonomy from prior prompt-stealing methods and build an automated stealing prompt generation agent. Starting from model-generated seed prompts, the framework expands attacks through scenario rationalization and structure injection while enforcing diversity via embedding-based filtering, yielding a reproducible pipeline for evaluating proprietary agent systems. We evaluate these attacks across commercial agent platforms and representative LLMs. Our results show that agent skills can often be extracted easily, posing a serious copyright risk. To mitigate this threat, we design defenses across the agent pipeline, focusing on input, inference, and output phase. Although these defenses substantially reduce leakage, the attack remains inexpensive and repeatable, and a single successful attempt is sufficient to compromise the protected skill. Overall, our findings suggest that these copyright risks remain largely overlooked across proprietary agent ecosystems, motivating stronger protection mechanisms.

cs.CR↗

arXiv2026

Efficient Near Field Beam Tracking via Thompson Sampling

The shift to the radiative near field region due to large antenna arrays necessitates beamforming that accounts for both angle and range, evolving mobility management into a joint angular range tracking challenge. Conventional schemes rely on rigid pilot payload structures with dedicated training slots, which interrupt data transmission and degrade spectral efficiency. To address this, we propose a pilot-free beam tracking framework leveraging Thompson sampling(TS). Within each sliding window, the user trajectory is modeled by local low-order polynomials in angle and range, and the motion parameters are estimated by maximum likelihood with uncertainty quantified via the Fisher information matrix. TS adaptively probes uncertain trajectory regions using beams that simultaneously serve as payload beams. Simulations demonstrate that the proposed framework maintains reliable connectivity while eliminating the overhead of dedicated pilot-based beam sweeping.

eess.SP↗

arXiv2026

Sensing-Assisted Secure Communication in MA-Aided ISAC: CRB Analysis and Robust Design

A core challenge in physical-layer security is the difficulty of obtaining the channel state information (CSI) of potential eavesdroppers. The inherent sensing functionality of integrated sensing and communication (ISAC) systems offers a promising solution by enabling the estimation of key parameters, such as the eavesdropper's angles of departure (AoDs). Capitalizing on this capability, we propose a sensing-assisted secure communication scheme for a movable antenna (MA)-aided ISAC system. The scheme comprises two stages: eavesdropper AoD sensing and secure communication. In the first stage, the base station (BS) optimizes the positions of its transmit and receive MAs to enhance sensing accuracy. We derive the closed-form Cramer-Rao bound (CRB) for the estimated AoDs to fundamentally characterize how MA positions influence the estimation uncertainty. In the second stage, the BS ensures secure communication by designing a robust beamforming vector that accounts for the AoD uncertainty region and by further optimizing the transmit MAs' positions to maximize the secrecy rate. To manage the end-to-end design, we formulate a joint optimization problem. This intractable non-convex problem is decomposed into two subproblems. For the first subproblem, we develop an alternating optimization (AO) algorithm to solve the CRB minimization problem. For the second subproblem, we solve the worst-case secrecy rate maximization problem using a method based on backward induction, convex hull construction, and AO. Finally, simulation results are provided to demonstrate the significant advantages of the proposed scheme compared to various benchmarks.

cs.IT↗

arXiv2026

Adaptive Sparse Group Lasso Penalized Quantile Regression via Dual ADMM

Sparse penalized quantile regression provides an effective framework for variable selection and robust estimation in high-dimensional data analysis. When ex planatory variables are organized into groups, achieving sparsity both within and between groups is essential. However, existing quantile regression methods often fail to meet this dual objective. To address this gap, we introduce the adaptive sparse group lasso penalized quantile regression, which integrates adaptive lasso and adaptive group lasso penalties. We optimize the model parameters via the alternating direction method of multipliers (ADMM) applied to the dual problem, and establish global convergence. Through extensive simulation studies and real data analyses, we demonstrate (i) the efficacy of the proposed method in achieving simultaneous within- and between-group sparsity, and (ii) the computational efficiency of our algorithm relative to existing alternatives.

stat.CO↗

arXiv2026

LKV: End-to-End Learning of Head-wise Budgets and Token Selection for LLM KV Cache Eviction

Long-context inference in Large Language Models (LLMs) is bottlenecked by the linear growth of Key-Value (KV) cache memory. Existing KV cache compression paradigms are fundamentally limited by heuristics: heuristic budgeting relies on statistical priors rather than task objectives, causing resource misallocation, while heuristic selection relies on coupled query-key interactions or static inductive biases (e.g., attention sinks). To address this limitation, we introduce LKV (Learned KV Eviction), which formulates KV compression as an end-to-end differentiable optimization problem. LKV integrates LKV-H to learn task-optimized global budgets, and LKV-T to derive intrinsic KV importance without materializing attention matrices. This design bypasses heuristic proxies, strictly aligning compression with task objectives. Extensive evaluations demonstrate that LKV achieves state-of-the-art performance on both LongBench and RULER benchmarks at high compression rates. In particular, on LongBench, LKV achieves near-lossless performance with only 15\% KV cache retention. Crucially, our analysis identifies learned budgeting as the dominant driver of fidelity, demonstrating that data-driven allocation is essential to overcome the limitations of hand-crafted heuristics.

cs.LG↗

arXiv2026

SentiAvatar: Towards Expressive and Interactive Digital Humans

We present SentiAvatar, a framework for building expressive interactive 3D digital humans, and use it to create SuSu, a virtual character that speaks, gestures, and emotes in real time. Achieving such a system remains challenging, as it requires jointly addressing three key problems: the lack of large-scale, high-quality multimodal data, robust semantic-to-motion mapping, and fine-grained frame-level motion-prosody synchronization. To solve these problems, first, we build SuSuInterActs (21K clips, 37 hours), a dialogue corpus captured via optical motion capture around a single character with synchronized speech, full-body motion, and facial expressions. Second, we pre-train a Motion Foundation Model on 200K+ motion sequences, equipping it with rich action priors that go well beyond the conversation. We then propose an audio-aware plan-then-infill architecture that decouples sentence-level semantic planning from frame-level prosody-driven interpolation, so that generated motions are both semantically appropriate and rhythmically aligned with speech. Experiments show that SentiAvatar achieves state-of-the-art on both SuSuInterActs (R@1 43.64%, nearly 2 times the best baseline) and BEATv2 (FGD 4.941, BC 8.078), producing 6s of output in 0.3s with unlimited multi-turn streaming. The source code, model, and dataset are available at https://sentiavatar.github.io.

cs.CV↗

arXiv2026

AGMA: Adaptive Gaussian Mixture Anchors for Prior-Guided Multimodal Human Trajectory Forecasting

Human trajectory forecasting requires capturing the multimodal nature of pedestrian behavior. However, existing approaches suffer from prior misalignment. Their learned or fixed priors often fail to capture the full distribution of plausible futures, limiting both prediction accuracy and diversity. We theoretically establish that prediction error is lower-bounded by prior quality, making prior modeling a key performance bottleneck. Guided by this insight, we propose AGMA (Adaptive Gaussian Mixture Anchors), which constructs expressive priors through two stages: extracting diverse behavioral patterns from training data and distilling them into a scene-adaptive global prior for inference. Extensive experiments on ETH-UCY, Stanford Drone, and JRDB datasets demonstrate that AGMA achieves state-of-the-art performance, confirming the critical role of high-quality priors in trajectory forecasting.

cs.CV↗