Search arXiv⌕ Search

arXiv subjects

Xin Zheng

Publications and source records attributed to Xin Zheng.

At least 19 recordsLinked to original sources
arXiv2026

Near-Logarithmic Inapproximability of Parameterized Set Cover

In the $γ$-gap $k$-SetCover problem, we are given a family $\mathcal{S}$ over a universe $U$, and the goal is to distinguish between the case where $\mathcal{S}$ contains $k$ sets covering $U$, and the case where at least $γk$ sets in $\mathcal{S}$ are needed to cover $U$. We show that, for every $k^{Ω(k^2)}\leρ\le n^{O(1)}$, it is W[1]-hard to approximate $k$-SetCover within ratio $Ω\big(\frac{\logρ}{k^2\log\logρ}\big)$ on instances with $n$ sets and universe size at most $ρ\log n$. We also show that under the randomized $k$-OV Hypothesis, there is no bounded-error randomized $f(k)n^{o(k)}$-time $o\big(\frac{\log N}{\log\log N}\big)$-approximation algorithm for $k$-SetCover. These results improve on the previous $(\log n)^{1/O(k)}$-inapproximability. For fixed $k$, this matches the polynomial-time $(\lnρ+O(1))$-approximation up to a $\log\logρ$ factor. We further establish a connection between small universes and large approximation gaps. We show that for every $1\leγ\le\min\big\{\frac{\log u}{\log\log N},\frac{\log u}{k^2\log\log u}\big\}$, exact $k$-SetCover with universe size $u^{Θ(1/γ)}$ is equivalent to $γ$-gap $k$-SetCover with universe size $u$ under randomized Turing reductions. This suggests a barrier to proving superconstant-gap W[2]-hardness: such a result would also make exact $k$-SetCover with universe size $n^{o(1)}$ W[2]-hard under randomized FPT Turing reductions. Finally, we analyze the VC dimensions of the set systems produced by our reduction. Write $d$ and $d^*$ for the primal and dual VC dimensions respectively. We show that for $γ=o(\frac{\log N}{\log\log N})$, $γ$-gap $k$-SetCover is W[1]-hard even when $d=O(k^{2}γ\log(kγ))$ and $d^*=O(kγ)$. The bound for $d^*$ matches the $O(d^*\log k)$-approximation algorithm up to a $k\log k$ factor.

cs.CC↗
arXiv2026

Dynamic Graph World Modelling: Learning Future Dynamics from Temporal Context

Dynamic graph learning models how entities and their relations evolve from historical observations to predict future graph states. Existing methods do not fully address dynamic graph world modelling, which requires learning transitions from temporal context to future latent graph states. This is difficult for three reasons: nodes and relations rely on different historical evidence, structural evolution is path-dependent, and future snapshots provide sparse and realization-specific targets. We propose Dynamic Graph World Modelling (DyG-WM), the first framework to jointly predict future node and relation representations from temporal context by answering three questions: what has evolved, how it evolved, and what will evolve next. To capture what has evolved, a historical node-relation context encoder summarizes node-neighborhood trajectories and pair-specific histories. To characterize how it evolved, a node-relation path-signature encoder captures ordered structural changes. To predict what will evolve next, a dual-level joint-embedding objective predicts target-snapshot node and relation representations without reconstructing the raw graph. Experiments on 12 link-prediction and two node-classification datasets demonstrate the effectiveness of DyG-WM across both tasks.

cs.LG↗
arXiv2026

Beyond In-Distribution Preservation: Recovering Generalization in Quantized VLAs via Vulnerability-Oriented Tuning

Post-training quantization has been shown to preserve VLA performance under standard evaluation conditions, but whether it preserves the full-precision model's robustness and generalization remains underexplored. In this study, we systematically study the robustness and generalization of post-quantized VLA policies under environmental disturbances. Empirical results show that quantized policies can become fragile to subtle environmental variations despite retaining comparable in-distribution performance. We further observe that action discrepancies are concentrated in a small subset of rollout states, while teacher guidance has opposite effects depending on discrepancy: it improves generalization at high-discrepancy states but can degrade it at low-discrepancy states. These findings reveal that effective post-quantization recovery requires selectively intervening on vulnerable states rather than globally distilling the student. We therefore propose Policy-Induced Vulnerability-Oriented Tuning (PIVOT-Q), a vulnerability-aware On-Policy Distillation (OPD) framework that selectively corrects vulnerable states encountered during quantized-student rollouts using the frozen full-precision policy as a teacher. PIVOT-Q identifies vulnerable states using discounted accumulated discrepancies over a short horizon, applies phase-balanced sparse supervision, and uses a Behavioral Anchor to prevent unnecessary changes. Experiments under seven LIBERO-Plus environmental variations demonstrate consistent recovery across multiple VLA backbones and quantization methods. Notably, PIVOT-Q consistently outperforms full-state distillation across all settings while using only 7.4% of its state-level distillation budget. Our code is available at https://github.com/ruanruan-andy/PIVOT-Q.

cs.LG↗
arXiv2026

Interpretable Hypergraph Learning via Neural Additive Models

Hypergraphs offer a natural framework for modeling networked data, where dependencies among entities are governed by higher-order interactions. While hypergraph learning methods such as hypergraph neural networks have demonstrated remarkable predictive performance, most existing approaches rely on black-box message-passing architectures, making it difficult to disentangle the contributions of node attributes and higher-order structural information. To address this challenge, we introduce the hypergraph neural additive network (HGNAN), an inherently interpretable framework for learning on hypergraph-structured data. HGNAN extends classical neural additive models to higher-order relational data by integrating feature-wise nonlinear decomposition with hypergraph-aware structural aggregation, enabling transparent prediction for both node- and hyperedge-level tasks. Extensive experiments on benchmark datasets demonstrate that HGNAN achieves performance comparable with state-of-the-art hypergraph learning methods while providing intrinsic and meaningful interpretability.

cs.LG↗
arXiv2026

An S-matrix Formalism for the Nonclassical Optical Response of Plasmonic Nanowires

This paper reports a computational scheme which predicts the light-matter interaction in multiple mesoscopic plasmonic nanowires (NWs), each of which includes concentric or eccentric constituent wires. Lying at the core of the proposed scheme is the S-Matrix of each individual cylindrical interface. The optical response of the materials inside and outside each interface can be described by three popular mesoscopic models: the Hydrodynamic Drude model (HDM) and its diffusive variant, namely the Generalized Nonlocal Optical Response (GNOR) model, as well as the Surface Response Model (SRM). To take into account the relative displacements between the cylindrical interfaces, the addition theorem is applied to construct the main equation behind the algorithm, where the possibility of longitudinal waves intrinsic in the HDM and GNOR models is also included. The optical responses calculated by the proposed S Matrix Method are compared with an in-house developed boundary element method (BEM) toolbox for a single gold NW and an eccentric dual-material NW, and excellent agreements are seen. The algorithm is further physically checked for a sodium nanolens trimer featuring large field focusing, and the effects of the HDM and the SRM are compared.

physics.comp-ph↗
arXiv2026

World-as-Graph: Relational World Modeling Through Latent Space Graphs

World models aim to learn representations of real-world environments and predict their future evolution. Recent object-centric world models have made expressive progress by representing visual scenes as sets of object-level latent states, but object-object relations are often captured only implicitly, which limits explicit relational and temporal structure modeling and object-centric dynamic memory modeling. To address such challenges, we propose World-As-Graph (WAG), a graph-based object-centric world model that introduces relational inductive bias into JEPA-style predictive representation learning. The proposed WAG contains two main modules: (1) Relation-aware structure induction, which constructs time-varying latent graphs from object-centric slots and designs relation-aware object masking policies to guide relational object representation learning in latent space; (2) Object-centric memory transition, which maintains and updates object-level dynamic states by combining relational information from neighboring objects with historical memory, enabling effective autoregressive future prediction. Extensive experiments on both visual reasoning and robotic manipulation tasks could demonstrate the superior performance of our proposed WAG.

cs.LG↗
arXiv2026

Frame Differential On-Policy Self-Distillation for Video Reasoning

Reinforcement learning (RL) has substantially improved the reasoning ability of multimodal language models through verifiable rewards and increasingly fine-grainedvisual or temporal credit assignment. In video reasoning, however, current RL methods typically train with a fixed sparse frame budget: increasing the number of frames makes autoregressive rollouts expensive, while too few frames may miss temporally localized events and fine-grained visual details. We present \textbf{Frame Differential On-Policy Self-Distillation (FD-OPSD)}, which transfers the useful evidence of dense frame observations to a sparse frame policy during RL training. FD-OPSD compares the policy's token level preferences for the same sampled response under sparse and dense views, and distills the resulting frame differential signal without an external teacher or dense autoregressive rollout. The method preserves sparse-frame rollouts and leaves inference unchanged. Across Qwen2.5-VL-7B and Qwen3-VL-4B on six video reasoning benchmarks, FD-OPSD yields higher overall average performance than the strongest corresponding GRPO, T-GRPO, or Video-KTR baselines across the 16, 32, and 64 frame evaluation settings. These results show that dense visual evidence can be transferred selectively during training through token level self-distillation while retaining sparse frame rollouts and unchanged inference.

cs.CV↗
arXiv2026

StreamRig: Exploiting Intra-Rig Geometry for Streaming Multi-Camera Odometry

Mobile robots and vehicles carry synchronized multi-camera rigs, yet many streaming 3D foundation models are designed for monocular input, leaving efficient use of rig geometry a challenge. We present StreamRig, a freeze-and-stream framework that builds causal streaming odometry for calibrated rigs on a frozen multi-view 3D foundation model. The frozen front-end jointly perceives the synchronized views using rig calibration. A Rig-Resampler compresses their features, a CausalBridge applies causal attention with a key-value cache, and a lightweight head regresses rig poses. A periodic re-anchoring protocol supports stable pose estimation over long sequences. Only these modules are trained, 74.6M parameters in total, with relative poses as the sole supervision. Our two-stage training strategy combines group relocalization pretraining with causal rig training to transfer the geometric priors of the frozen front-end and the alignment ability of the pretrained modules to streaming odometry. We evaluate on NCLT, TartanGround, KITTI-360, and our self-collected humanoid-robot dataset ZJH, where training uses only simulation and real-world evaluation is zero-shot. Across all four datasets, StreamRig achieves lower translation and rotation drift than the evaluated non-oracle monocular streaming and rig-aware offline models, while maintaining low inference cost. Ablations and controlled camera-count experiments identify the sources of these gains. We further examine how longer training windows affect inference over longer horizons. Code has been released at https://github.com/WeiYuFei0217/StreamRig.

cs.CV↗
arXiv2026

OntologyAligner: Ontology-Aligned Retrieval and Hierarchy-Guided Large Language Model Reranking for Biomedical Ontology Normalization

Biomedical ontology normalization maps free-text expressions to standardized concepts, enabling consistent integration and analysis of biomedical data. This task remains challenging because lexical variation and subtle distinctions among hierarchically related concepts can obscure concept boundaries. We present OntologyAligner, a three-stage framework that combines ontology-aligned retrieval, large language model candidate reranking, and selective hierarchy-guided refinement. We also construct PhenoNormBench, a unified benchmark comprising 13,390 samples from seven Human Phenotype Ontology datasets. OntologyAligner achieved state-of-the-art performance on HPO normalization, with 88.78% Macro Top-1 Accuracy and 86.75% Micro Top-1 Accuracy, exceeding the strongest baseline by 4.85 and 5.07 percentage points, respectively. Ablation analyses showed complementary contributions from all three stages, and sensitivity analyses demonstrated stability across candidate-set sizes and model backbones. Applications to MONDO, MEDIC, and NCBITaxon further established portability to other ontologies. OntologyAligner offers a generalizable framework for accurate mapping of biomedical text to structured ontology concepts. PhenoNormBench and the code are publicly available at https://github.com/zhelishisongjie/OntologyAligner.

cs.AI↗
arXiv2026

Temporal Memory-Aware Online Test-Time Adaptation on Dynamic Graphs

Test-time adaptation (TTA) on graphs aims to adapt a graph neural network (GNN) that is well-trained on the training graph to the test graph, which involves potential distribution shifts that may harm model generalization and test-time inference. While recent efforts have investigated TTA on static graphs, there is still a research gap on dynamic graphs learned with dynamic GNN (DGNN) models, where both structural connectivity and node semantics evolve continuously over time. This makes adapting a DGNN model for reliable test-time performance substantially challenging. To fill this gap, in this work, we propose a novel framework of temporal memory-aware Online Test-Time Adaptation on Dynamic Graphs, named DGOTTA, to effectively adapt well-trained DGNNs during test time. Specifically, the proposed DGOTTA contains three modules: (1) temporal-aware augmentation, to extend the diversity of test dynamic graphs for addressing complex temporal and spatial shifts; (2) memory-aware model prediction, to alleviate catastrophic forgetting; (3) consistency-guided online adaptation, to enforce temporal alignment and memory smoothness. Extensive experiments on three real-world datasets and four DGNN backbones demonstrate that DGOTTA significantly improves generalization under diverse distribution shifts and multiple model architectures.

cs.LG↗
arXiv2026

OpenMUSTANC (MUltiple Scattering Theory At Nanoplasmonic Cavities): A MATLAB toolbox for the simulation of Plasmonic Sphere Aggregates

Mesoscopic physical models, including the Hydrodynamic Drude Model (HDM), the Generalized Nonlocal Optical Response (GNOR) Model, and the Surface Response Model (SRM), have been proposed to investigate nonlocal effects in nanometric structures. The combination of classical electromagnetism with these mesoscopic material models calls for new computational electromagnetic (CEM) algorithms, or update of conventional ones, in what is termed computational mesoscopic electromagnetics (CMEM). In this work, we present a MATLAB toolbox for the simulation of multiple spherical interfaces with arbitrary relative positions, with the incorporation of the aforementioned mesoscopic models. The method exploits vector spherical wave functions to properly express electric and magnetic fields, an S matrix formulation for the efficient treatment of incident and scattered fields at spherical interfaces, and a translation matrix to deal with propagating waves with different expansion centers. Excitation sources can be chosen among arbitrarily polarized plane waves, dipoles and electron beams. For the post-processing part, the calculation of cross sections and far/near-field mapping; fluorescence enhancement, Purcell factor and quantum yield; and cathodoluminescence and electron energy-loss probability is implemented. The toolbox is built in a modular manner, and each part (routine) has its own important functionality. This paper provides a full explanation of the proposed highly efficient and general toolbox, and a detailed guideline for researchers in the nanoplasmonics community.

physics.comp-ph↗
arXiv2026

SeMoCo: A Semantic-First Motion Codec for Motion Language Modeling

Discrete motion representations have substantially advanced autoregressive text-to-motion generation. However, most motion tokenizers are optimized for reconstruction and do not explicitly allocate capacity according to semantic role. Action-level meaning and fine-grained kinematic detail must therefore be encoded through the same reconstruction-driven hierarchy. We introduce SeMoCo, a semantic-first motion codec, together with a dual-axis motion generator for language-conditioned motion generation. Each motion token contains one semantic token and a residual sequence of kinematic tokens. The generator models semantic progression across time and autoregressively refines the residual entries. We also construct $Ω$-MotionVerse, a large-scale, multi-source human-motion dataset unified under the SOMA representation. Across the reported comparisons, SeMoCo achieves the best reconstruction accuracy among the compared codecs, while strong text-to-motion results demonstrate the effectiveness of its motion tokens for downstream generation.

cs.CV↗
arXiv2026

Explain Before You Answer: A Survey on Compositional Visual Reasoning

Compositional visual reasoning has emerged as a key research frontier in multimodal AI, aiming to endow machines with the human-like ability to decompose visual scenes, ground intermediate concepts, and perform multi-step logical inference. While early surveys focus on monolithic vision-language models or general multimodal reasoning, a dedicated synthesis of the rapidly expanding compositional visual reasoning literature is still missing. We fill this gap with a comprehensive survey spanning 2023 to 2025 that systematically reviews 260+ papers from top venues (CVPR, ICCV, NeurIPS, ICML, ACL, etc.). We first formalize core definitions and describe why compositional approaches offer advantages in cognitive alignment, semantic fidelity, robustness, interpretability, and data efficiency. Next, we trace a five-stage paradigm shift: from prompt-enhanced language-centric pipelines, through tool-enhanced LLMs and tool-enhanced VLMs, to recently minted chain-of-thought reasoning and unified agentic VLMs, highlighting their architectural designs, strengths, and limitations. We then catalog 60+ benchmarks and corresponding metrics that probe compositional visual reasoning along dimensions such as grounding accuracy, chain-of-thought faithfulness, and high-resolution perception. Drawing on these analyses, we distill key insights, identify open challenges (e.g., limitations of LLM-based reasoning, hallucination, a bias toward deductive reasoning, scalable supervision, tool integration, and benchmark limitations), and outline future directions, including world-model integration, human-AI collaborative reasoning, and richer evaluation protocols. By offering a unified taxonomy, historical roadmap, and critical outlook, this survey aims to serve as a foundational reference and inspire the next generation of compositional visual reasoning research.

cs.CV↗
arXiv2026

$\ell_1$-Based Adaptive Identification under Quantized Observations with Applications

Quantized observations are ubiquitous in a wide range of applications across engineering and the social sciences, and algorithms based on the $\ell_1$-norm are well recognized for their robustness to outliers compared with their $\ell_2$-based counterparts. Nevertheless, adaptive identification methods that integrate quantized observations with $\ell_1$-optimization remain largely underexplored. Motivated by this gap, we develop a novel $\ell_1$-based adaptive identification algorithm specifically designed for quantized observations. Without relying on the traditional persistent excitation condition, we establish global convergence of the parameter estimates to their true values and show that the average regret asymptotically vanishes as the data size increases. Finally, we apply our new identification algorithm to a judicial sentencing problem using real-world data, which demonstrates its superior performance and practical significance.

eess.SY↗
arXiv2026

On the Approximability of Parameterized Minimum Monotone Satisfying Assignment

The parameterized Minimum Monotone Satisfying Assignment ($k$-MMSA) problem asks whether a monotone Boolean circuit admits a satisfying assignment of Hamming weight at most $k$. The MMSA hierarchy is defined by allowing a bounded number of alternations between AND and OR gates in the circuit. While the polynomial-time approximability of the MMSA hierarchy has been studied extensively, much less is known in the parameterized setting. In particular, $k$-MMSA$_2$ is the well-known $k$-SetCover problem, whose parameterized inapproximability lies in the $\text{polylog}(n)$ regime. In contrast, $k$-MMSA$_4$ captures $k$-MinLabel, for which known lower bounds give $\text{poly}(n)$ inapproximability. Sandwiched by $k$-MMSA$_2$ and $k$-MMSA$_4$, the inapproximability of $k$-MMSA$_3$ remained comparatively unexplored. In this paper, we give an FPT-time $O(2^k \log n)$-approximation algorithm for $k$-MMSA$_3$, suggesting that in the fixed-parameter regime, the third level of MMSA remains surprisingly close to the second level. Complementing this algorithm, we also give an FPT-time gap-preserving reduction from $k$-MMSA$_3$ to $k$-MMSA$_2$. Thus, stronger inapproximability for $k$-MMSA$_3$ would imply new hardness for $k$-MMSA$_2$, potentially offering a route around the current barriers for the latter problem. Revisiting Marx's reduction from $k$-MMSA$_t$ to gap $k$-MMSA$_{t+2}$, we also show that $k$-MMSA$_4$ admits no $n^{o(1)}$-factor FPT approximation unless W[2]=FPT, and no $n^{O(1/k)}$-factor approximation running in $n^{o(k)}$ time under ETH. These results separate the parameterized approximability behavior of the third and fourth levels and clarify where stronger inapproximability enters the $k$-MMSA hierarchy.

cs.CC↗
arXiv2026

SoK: Attack and Defense Landscape of Mobile On-device AI Systems

Mobile on-device AI (MoAI) systems that integrate locally deployed AI models with conventional mobile software components are emerging as a key paradigm for delivering intelligent functionality directly on end-user devices. By moving inference from remote cloud services to the local mobile environment, such systems enable privacy-preserving, low-latency, and offline-capable AI functionality, yet introduce new security risks arising from the local storage of AI models. This paper presents the first comprehensive systematization of knowledge on MoAI security, covering security pillars, attack landscape, and defense landscape of MoAI systems. We further identify unresolved gaps in current attack and defense research and point to promising directions for future research in this emerging area. Our work establishes the first systematic framework for understanding the attack and defense landscapes of MoAI systems, serving as a foundation for building secure MoAI systems and advancing research in this critical domain. Companion resources are available at https://github.com/Jinxhy/Awesome-MoAI-Security.

cs.CR↗
arXiv2026

Analysis of Adam Algorithms for Stochastic Dynamic Systems

The adaptive moment estimation algorithm, known as Adam, is widely used in modern machine learning, owing to its low per-iteration complexity and strong empirical performance. Despite its prevalent use, the theoretical foundation of Adam remains largely unexplored for time-varying and nonstationary systems. In fact, the existing theoretical analyses of Adam-type algorithms are primarily concerned with time-invariant model parameters and explicitly or implicitly rely on independent and identically distributed (i.i.d.) data assumptions, under which the learning taskcan be formulated as minimizing a fixed expected objective with a static minimizer. However, such assumptions are often violated in time-varying and nonstationary systems, thereby calling for a theoretical investigation beyond the conventional yet idealized i.i.d. setting. The main objective of this paper is to solve this challenging problem by establishing a general theory of Adam for time-varying and nonstationary stochastic systems. We will introduce some new techniques for analyzing the products of nonstationary and dependent random matrices induced by Adam's coupled first- and second-moment recursions, and will construct a new stochastic Lyapunov function that blends these two moment dynamics. Under a stochastic excitation condition that allows nonstationary and dependent data, we will derive both parameter tracking and output prediction error bounds explicitly, quantifying the effects of stepsize, first- and second-momentum parameters, gradient noise and parameter drift. These bounds not only provide guarantees for Adam performance, but also provide guidelines for hyperparameter selection. Experiments on both synthetic and real-world data validate our theory and design guidelines.

cs.LG↗
arXiv2026

Broadband on-chip Half Maxwell Fisheye

In this article, we report on the design and the experimental evidence of a Half Maxwell Fish Eye (HMFE), for Silicon Photonics and working at telecommunication wavelength. It is designed by implementing a Graded Photonic Crystal operating in the non-resonant metamaterial regime. The results of 3D Finite-Difference Time-Domain simulations (FDTD) show an excellent broadband focusing capacity. It has been urther fabricated via the Silicon On Insulator (SOI) platform for its compatibility with CMOS technology. Experimentally, its performances are firstly investigated by the means of a fan-shaped set output waveguides. Next, Scanning Near-Field Optical Microscopy (SNOM) characterisation confirms the wavefront curving inside the HMFE lens. Quantitative analysis of the SNOM results demonstrates its excellent focusing performances: the Full Width Half Maximum (FWHM) is $0.466λ_0$ at $λ_0=1550$nm, while the thickness of the lens is $3.18λ_0$.

physics.optics↗