Search arXiv⌕ Search

arXiv subjects

Zhe Wang

Publications and source records attributed to Zhe Wang.

At least 73 records · Page 4Linked to original sources
arXiv2026

SD-MAR: Multi-image Analytical Reasoning via Synthetic Data and Reinforcement Learning

Vision Language Models (VLMs) demonstrate strong perceptual abilities but remain limited in tasks requiring analytical reasoning across multiple visual states, such as multi-image comparison, change detection, and multi-step visual inference. These capabilities are critical for real-world multimodal applications where reasoning must be grounded in systematic differences between visual contexts. However, existing benchmarks rarely require both explicit visual comparison and analytical reasoning, leaving this capability underexplored. To address this gap, we introduce SD-MAR (Synthetic Data for Multi-image Analytical Reasoning), a framework for training and evaluating VLMs on multi-image analytical reasoning. SD-MAR constructs paired visual scenarios through controlled perturbations and generates reasoning tasks spanning semantic change attribution and quantitative comparison. We further train VLMs using GRPO-lite with Backward Discounted Allocation (BDA), a reinforcement learning approach that removes KL regularization to encourage stronger policy optimization while allocating greater credit to the later reasoning steps where analytical conclusions are formed. Experiments on Qwen2.5-VL-7B and InternVL3-8B show that GRPO-lite fine-tuning on SD-MAR improves in-domain accuracy by up to 36.95%, with Qwen2.5-VL-7B outperforming GPT-4.1 on the SD-MAR benchmark. Importantly, out-of-domain generalization is preserved or improved: performance remains within 1% on MME, MMMU-Pro, and MathVista, while improving by up to 4% on MMBench. LLM-as-judge evaluation further demonstrates consistent improvements in logical coherence and explanation quality across both models.

cs.CV↗
arXiv2026

Logarithmic corrections to bulk and surface criticality in a three-dimensional quantum Heisenberg antiferromagnet

At the bulk upper critical dimension, marginally irrelevant interactions generate multiplicative logarithmic corrections to mean-field scaling. While these corrections are well understood for bulk observables, their consequences for boundary criticality, particularly for finite-size scaling, remain much less explored. Here we combine large-scale quantum Monte Carlo simulations with boundary renormalization-group analysis to study a (3 + 1)D O(3) quantum critical point. After verifying the known logarithmically modified bulk finite-size scaling, including the correlation-length scaling governed by the logarithmic finite-size exponent \hat{\coppa}, we tune the surface coupling to identify ordinary, special, and extraordinary boundary regimes. For the ordinary and special transitions, we derive logarithmic correction exponents and \hat{\coppa}-dependent finite-size scaling forms for boundary correlations, including results that have not been systematically established before. These predictions are quantitatively supported by Monte Carlo data. In the extraordinary regime, we find long-range surface magnetic order and a logarithmically enhanced surface-bulk correlation.

cond-mat.str-el↗
arXiv2026

PrismAD: Decoupled Planning via Semantic Mixture-of-Planners for End-to-End Autonomous Driving

This letter presents PrismAD, a decoupled end-to-end autonomous driving framework based on a Semantic Mixture-of-Planners. Existing planners usually aggregate heterogeneous scene tokens into a coupled representation space, forcing a single planning branch to jointly model agent interaction, road geometry, and driving intention. Such coupling may weaken factor-specific reasoning and obscure the contribution of different planning cues. To address this limitation, PrismAD partitions scene tokens into interaction, geometry, and intent groups, and assigns them to independent planning experts with the same architecture but separate parameters. Each expert learns a specialized motion-planning representation, while a semantics-aware router adaptively aggregates expert predictions with separate routing weights for motion prediction and ego planning. Sparse top-$K$ activation with noisy gating is further introduced to improve routing robustness and reduce unnecessary expert computation. Extensive experiments on the nuScenes open-loop dataset and NeuroNCAP closed-loop benchmark demonstrate that PrismAD exhibits competitive performance. Our code will be released soon.

cs.RO↗
arXiv2026

Reinforcing the Generation Order of Multimodal Masked Diffusion Models

Diffusion Language Models (DLMs) have recently achieved substantial progress in natural language generation tasks. Recent research demonstrates that adaptive token generation ordering can significantly improve performance in mathematical reasoning and code synthesis applications. In this work, we investigate the optimization of generation order for both text-to-image synthesis and multimodal understanding. We first establish that, unlike structured problems in language generation such as Sudoku puzzles, model logits alone are insufficient for determining optimal generation sequences in text-to-image generation and multimodal understanding. To address this challenge, we introduce a learnable control module trained via Group Relative Policy Optimization (GRPO) to determine the generation order. Our results demonstrate that learning this control block substantially improves both text-to-image alignment and multimodal understanding in DLMs. In particular, it enhances the model's ability to capture fine-grained spatial relationships in generated images while also strengthening performance on multimodal reasoning and comprehension tasks. We evaluate our framework on GenEval, an object-focused benchmark for text-to-image alignment, where it achieves 4.08% relative improvements. In addition, experiments on VLMEvalKit confirm 4.85% relative improvements in multimodal understanding, highlighting the broad effectiveness of our approach.

cs.LG↗
arXiv2026

Synthetic-to-Real Translation for Class-Agnostic Motion Prediction

Motion understanding is critical for ensuring safety and robustness in autonomous driving systems, driving increasing interest in motion prediction. A key challenge in this domain is the high cost associated with acquiring real-world motion labels. It is therefore ideal if we could transfer motion knowledge from synthetic data to real data. In this context, we explore the potential of synthetic-to-real translation for motion prediction (SRMP). However, the most used naive motion regression methods are notably sensitive to the synthetic-to-real domain shift, resulting in unreliable knowledge translation. To address this, we propose a novel approach integrating a motion knowledge translation framework with two key components: (1) objectness-aware motion prediction, which explicitly models the joint distribution of motion patterns and objectness priors to improve domain-invariant feature learning, and (2) objectness-aided motion enhancement, a motion label refinement mechanism that leverages learned objectness priors to filter motion noise. Furthermore, we present a physically-based pipeline for generating Motion4D, the first synthetic 4D LiDAR dataset tailored for SRMP research, addressing the lack of synthetic motion datasets. Experimental results demonstrate that our approach effectively bridges the domain gaps and yields superior performance on real scenes.

cs.CV↗
arXiv2026

Unexpected Far-Near-Far Transition in Mobile Near Field Terahertz Communications

At THz frequencies, the radiative near-field distance can be sufficiently large to matter in real deployments. Existing near-field formulas are often understood in a simple way: as the link distance decreases, the propagation regime is expected to change only once, i.e., from far field to near field. This paper shows that this intuition can fail for an elevated access point with downward tilt serving a ground user moving along the ground. Along such a path, the link distance and the viewing angle change together, so the near-field to far-field transition may take place more than once, creating an unexpected far-near-far transition. In this paper, we derive analytical conditions for when this transition occurs for tilted ULA-to-point and UPA-to-point scenarios and compute the corresponding transition point(s) on the ground. Numerical results validate the analysis and further show that this behavior depends strongly on the deployment geometry and can also arise at lower frequencies.

eess.SP↗
arXiv2026

ATCCaps: A Call-Sign-Aware Speech Dataset for Air Traffic Control Recognition

Call signs are safety-critical entities in air traffic control (ATC) communications because they identify the target aircraft of each spoken instruction. This paper presents ATCCaps, a call-sign-aware ATC speech dataset with caption-level audio-text supervision. Built from real ATC radiotelephony recordings, ATCCaps contains 202.94 hours of curated audio, 170,385 utterances, and 922 unique normalized call signs. The construction pipeline combines confidence-aware transcript parsing, ADS-B-derived call-sign metadata, call-sign normalization, rule-based quality filtering, and LLM-assisted caption generation. Each retained sample is paired with transcript descriptions, call-sign descriptions, and ATC-style captions, supporting ASR evaluation, call-sign matching, and call-sign-aware audio-text retrieval. We further characterize ATCCaps through split statistics, call-sign coverage, seen/unseen call-sign analysis, filtering audits, and caption quality evaluation. The evaluation subset is derived from the human-annotated ATCO2-test-set, enabling reference evaluation with manual transcripts. Results show that ATCCaps provides scalable audio-grounded call-sign supervision, while caption analysis highlights the need to explicitly validate call-sign and numeric fidelity. Reference ASR and CLAP-based baselines demonstrate the usability of ATCCaps for call-sign-aware ATC speech modeling.

cs.SD↗
arXiv2026

Meromorphic maps from ${\Bbb C}^{p}$ into semi-abelian varieties and general projective varieties

In 1974, W. Stoll proposed a method of studying holomorphic functions of several complex variables by reducing them to one variable through fiber integration. In this paper, we use this method to extend some important Nevanlinna-type results for holomorphic curves into projective varieties to meromorphic maps from ${\Bbb C}^{p}$ to projective varieties. This includes Bloch's theorem and Noguchi-Winkelmann-Yamanoi's Second Main Theorem for holomorphic maps into semi-abelian varieties intersecting an effective divisor, as well as Huynh-Vu-Xie's Second Main Theorem for meromorphic maps into projective space intersecting with a generic hypersurface with sufficiently high degree.

math.CV↗
arXiv2026

Analysis of $J/ψ$ and $ψ(2S)$ Charmonium Production in Ultraperipheral Lead-Lead and Proton-Lead Collisions at LHC Energies

The two gluon exchange model serves as a key framework for describing the photoproduction of heavy vector mesons, and the photoproduction cross sections derived from it provide essential input for studies of Ultra-Peripheral Collisions (UPCs). Building on the model successful description of $J/ψ$ and $ψ(2S)$ photoproduction in previous work, we use the STARlight program to systematically investigate charmonium UPCs in PbPb and $p$Pb collisions at LHC energies. To reduce discrepancies between theoretical predictions and experimental data for rapidity and transverse momentum distributions in PbPb collisions, a phenomenological suppression factor is introduced to correct the theoretical results. We find that, in the TeV energy range, the corrected predictions agree well with experimental data within uncertainties and successfully reproduce the characteristic double-peak structure in rapidity distributions. In contrast, no significant suppression is observed in $p$Pb UPCs, which reflects the asymmetric photon fluxes and the dominant contribution from the photoproduction interaction branch. The transverse momentum distributions from STARlight simulations also match the diffractive pattern of coherent production seen experimentally, although the overall yield remains slightly overpredicted. This work further validates the applicability of photoproduction cross sections from the two gluon exchange model for charmonium UPC studies, and offers valuable phenomenological guidance for future experimental design and data analysis in UPC measurements of charmonium production.

hep-ph↗
arXiv2026

Room-temperature tuning and probing of Fermi polarons in atomically thin semiconductors on a plasmonic metasurface

The Fermi polaron, arising from interactions between a mobile impurity and a degenerate Fermi sea, is a many-body quasiparticle that provides a sensitive probe of strongly correlated electronic phases in atomically thin semiconductors. In doped transition-metal dichalcogenides, the attractive and repulsive polaron branches are well established in monolayers. However, extending active control and quantitative, branch-resolved probing to stacked geometries has remained elusive because spectral quenching and weak optical contrast restrict access to Fermi polaron signatures. Here, we integrate electron-doped WS$_2$ flakes from monolayer to quadrilayer with a strain-tunable plasmonic metasurface, enabling high-contrast scattering readout at room temperature through coupling between Fermi polaron resonances and surface plasmons. This platform enables quantitative extraction of polaron branch spectral weights and coupling strengths across different layer numbers. We uncover a systematic thickness dependence of the spectral-weight distribution and demonstrate continuous and fully reversible spectral-weight transfer between attractive and repulsive branches in bilayers and quadrilayers, with near-complete transfer achieved in bilayers. By identifying layer number and strain as complementary control parameters for Fermi polarons, our results establish metasurface-enabled scattering spectroscopy as a practical route to resolve and manipulate many-body resonances in stacked van der Waals semiconductors, bridging idealized monolayer polaron physics and device-relevant architectures.

cond-mat.mes-hall↗
arXiv2026

Variational Learning for Insertion-based Generation

Non-monotonic sequence generation methods, such as masked diffusion models, provide a flexible alternative to left-to-right autoregressive modeling by allowing tokens to be generated in non-fixed and prescribed orders. Despite their practical advantages, most existing non-monotonic models are order-agnostic and rely on a fixed-length grid, limiting their ability to support variable-length generation and adaptive insertion order. In this work, we introduce a probabilistic framework for learning insertion order in variable-length insertion models. We formalize a bijective correspondence between insertion trajectories and permutations, which enables an exact reparameterization of the data likelihood as a sum over permutations. Building on this result, we propose the Insertion Process (IP), a stochastic generative model that jointly learns where to insert, what to insert, and when to terminate, trained via permutation-based variational inference. Unlike prior fixed-canvas approaches, IP natively supports variable-length generation and learns data-driven preferences over insertion orders. Experiments on goal-conditioned planning and molecular string generation demonstrate that learning insertion order improves both modeling quality and generalization in domains without a canonical left-to-right structure.

cs.LG↗
arXiv2026

A quantitative approach to flowing supercooled liquids: From microscopic heterogeneities to rheology

Soft glassy materials display rich and complex flow behaviors across both macroscopic and molecular scales, and a fundamental understanding of these phenomena remains an outstanding challenge. Here, we propose a theoretical model for the flow of supercooled liquids -- a typical class of glassy fluids -- based on a two-state paradigm that conceptualizes the flow as a dynamic coexistence of transient solid-like and liquid-like regions. The model rests on two essential physical ingredients: a correlation length that captures medium-range structural order, and a localized elasticity-mediated interaction that restricts stress propagation within solid-like regions. Remarkably, with all parameters determined solely from equilibrium state, the model quantitatively reproduces rheological responses -- including both steady-state and start-up shear -- for a broad range of shear rates. Furthermore, it simultaneously captures the evolution of molecular dynamic heterogeneity. This dual success -- spanning macroscopic rheology and microscopic spatiotemporal fluctuations -- underscores the pivotal role of structural and dynamic heterogeneities in governing the rheological response. Moreover, it provides a direct understanding of how the flow behaviors of a supercooled liquid are embedded in its equilibrium properties.

cond-mat.soft↗
arXiv2026

Dynamic XR Rendering Offloading Based on Feature-Based Quality Assessment

Extended Reality (XR) applications demand intensive computation and low latency, especially for real-time rendering tasks. In this letter, we present an edge-aided XR rendering testbed that dynamically offloads rendering workloads between the XR client and the edge server built upon network conditions and latency constraints. The testbed integrates a Microsoft HoloLens 2 headset, a GPU-enabled edge server, and a customized remote rendering toolkit based on the HOLO Stream SDK, enabling seamless switching between local and edge rendering modes in real time. To overcome the limitations of pixel-level quality metrics under head movements and asynchronous frame arrivals, we propose a perceptual evaluation metric based on deep feature embeddings and cosine similarity, which remains robust to spatial and temporal misalignments. Furthermore, we design a contextual bandit learning controller to adapt rendering placement decisions in real time by jointly optimizing perceptual quality and latency. Experimental results demonstrate the feasibility and performance of our testbed, validating its effectiveness in delivering high-quality and interactive XR experiences.

eess.IV↗
arXiv2026

Mountain Muography for China Jinping Underground Laboratory

The China Jinping Underground Laboratory (CJPL), located $\sim 2,400$~m beneath Jinping Mountain, is one of the world's deepest and largest ($\sim 300{,}000~\mathrm{m}^3$) underground facilities, hosting dark matter, nuclear astrophysics, and neutrino experiments. We report the first muon radiography (muography) conducted at this extraordinary depth. Cosmic muons detected by a one-ton prototype developed for the Jinping Neutrino Experiment were used to perform non-invasive subsurface density mapping over a 3~km lateral range. The 1.3~m diameter detector provides nearly isotropic acceptance and an angular resolution of $\sim 4.5^\circ$. By correlating the predicted surface muon flux distributions with the underground measurements, we reconstruct a directional opacity map that constrains the density structure of the overburden and shows excellent agreement with satellite-derived terrain models. This work demonstrates the feasibility of muography at extreme depths with kilometer-scale overburden and establishes a robust methodology for future geophysical applications and large-scale facilities, such as the full Jinping Neutrino Experiment. Based on this validated overburden model, we further predict the total muon fluxes for the eight experimental halls in CJPL-II, providing essential input for their physics programs.

hep-ex↗
arXiv2026

A Pathology Foundation Model for Gastric Cancer with Real-World Validation

Gastric cancer remains a major cause of cancer mortality, yet its histological and molecular heterogeneity complicates diagnosis and risk stratification. General-purpose pathology foundation models (PFMs) often plateau on fine-grained endpoints central to gastric cancer care, and few have undergone rigorous prospective validation or clinical reader studies. We present GRACE, a Gastric-specific foundation model for Real-world Assessment and Clinical dEcision support. GRACE was developed from multicenter gastric pathology datasets totaling 48,364 primarily HE-stained whole-slide images from 37,493 patients. When evaluated on 28 clinically relevant tasks, GRACE consistently outperformed representative pancancer PFMs, achieving a macro-AUC of 0.9188, with strong performance for precancerous lesion diagnosis (macro-AUC 0.9322), tumor histopathological assessment (macro-AUC 0.9119), molecular profiling (macro-AUC 0.8682), and prognostic prediction. Beyond benchmarking, GRACE's translational value was substantiated through a rigorous evidence chain. Under safety-gated criteria requiring 100% NPV for rule-out and 100% PPV for rule-in, GRACE streamlined review for up to 69.6% of malignancy-diagnosis cases and triaged 46.8% of MMR-IHC follow-up requests. This translational feasibility was further strengthened by a randomized crossover reader study of pathologist-AI collaboration. With GRACE assistance, diagnostic accuracy improved from 82.0% to 89.9%, yielding nearly twofold higher adjusted odds of a correct diagnosis (OR 1.987) alongside concurrent gains in sensitivity and specificity. AI assistance also reduced diagnostic time by 14.9%, elevated diagnostic confidence by 9.0%, and markedly improved inter-rater agreement. When calibrated to maintain non-inferior performance to senior pathologists, the AI-assisted workflow could triage 60.7% of atrophy and 82.7% of intestinal metaplasia cases.

cs.CV↗
arXiv2026

Sparse Activation for Sustainable Cell-Free Massive MIMO Networks: Less is More

Motivated by the vision of making sixth-generation (6G) networks sustainable, we study the sparse antenna/array activation problems in uplink cell-free massive multiple-input multiple-output (CF mMIMO) networks. We first develop an antenna-level optimal bilinear equalizer (OBE) weighting framework, in which each access point-user equipment (AP-UE) pair is assigned a matrix-valued long-term weight to shape the contribution of individual antenna elements, thereby generalizing the conventional large-scale fading decoding (LSFD) strategy from scalar coefficients to antenna-element-aware weighting. Building on this structure, we formulate sparse antenna activation as structured sparsity-inducing mean square error (MSE) minimization problems, and design four activation schemes at two granularities: antenna-level and array-level, each with UE-specific and network-wide (all-UEs) variants. The resulting convex problems are solved efficiently via the proximal method with closed-form group-wise updates, while the network-wide schemes are modeled through hierarchical sparsity and handled by a tree-structured proximal operator. Numerical results under correlated Rician channels and a detailed power consumption model demonstrate that the OBE weighting scheme consistently improves spectral efficiency over the LSFD, with gains increasing with the number of antennas. Meanwhile, the studied sparse activation schemes can achieve substantial energy efficiency improvement and power reduction with controllable spectral efficiency loss.

cs.IT↗
arXiv2026

Spatial Transcriptomics-Guided Alignment Enhances Molecular Profiling in Pathology Foundation Model

Comprehensive molecular profiling is essential for modern precision oncology but remains hindered by prohibitive costs, specimen exhaustion, and protracted turnaround times. While pathology foundation models (PFMs) have demonstrated potential for inferring molecular phenotypes from routine hematoxylin and eosin (H&E) whole-slide images (WSIs), current architectures primarily rely on vision-centric self-supervised learning or vision-language alignment, lacking the spatially resolved molecular supervision required to connect subtle morphological features with underlying genomic alterations. Spatial transcriptomics (ST) emerges as a transformative technology that enables transcriptomic quantification within intact tissue sections, thereby preserving the precise spatial link between histology and molecular profiles. In this study, we present a Spatial Transcriptomics-guided Alignment framework for Molecular Profiling (STAMP), which endows PFMs with intrinsic molecular awareness. To support this paradigm, we curated HumanST-1k, a human ST dataset spanning diverse anatomical organs and sequencing platforms. This atlas yields 1.8 million pairs of H&E patches and corresponding transcriptomic profiles, providing a corpus that links histological structures with their molecular states. To mitigate the technical noise inherent to raw transcriptomics, STAMP applies a pathway-informed alignment strategy that aggregates transcriptomic data into biologically functional pathways, which are subsequently integrated into PFMs via parameter-efficient fine-tuning. This alignment enriches the representation space of PFMs and unlocks their capacity to resolve sub-visual molecular signatures. The clinical utility of these augmented representations was validated through a multi-tier evaluation framework.

cs.LG↗
arXiv2026

EvalVerse: Pipeline-Aware and Expert-Calibrated Benchmarking for Professional Cinematic Video Generation

The rapid evolution of generative video foundation models has propelled the field toward professional-grade cinematic synthesis. To achieve such demanding quality, the community transitions towards Reinforcement Learning (RL) and agentic workflows. However, reliable evaluation has emerged as a critical bottleneck. Existing benchmarks predominantly evaluate ''whether it is right'' (basic prompt-following) while fundamentally neglecting ''whether it is good'' (cinematic quality, acting, and aesthetics). Furthermore, current automated metrics lack the domain-specific rigor required to provide trustworthy signals, creating a severe credibility gap between human aesthetic perception and machine scoring. To bridge this gap, we introduce EvalVerse, a comprehensive, pipeline-aware, and expert-calibrated evaluation framework. We treat video generation assessment not merely as an engineering task, but as a core scientific problem: the systematic digitization of subjective cinematic expertise. First, we organize domain knowledge into an evaluation taxonomy aligned with the professional filmmaking workflow (pre-production, production, and post-production). Second, we distill human expert judgments into a curated dataset with large-scale human annotations. Third, we inject this knowledge into Vision-Language Models (VLMs) through an expert-calibrated fine-tuning strategy, enabling the VLM to perform explicit Chain-of-Thought reasoning. Compared to previous works, EvalVerse not only retains compatibility with foundational ''rightness'' metrics, but also significantly expands the criteria to ''goodness'' and broaden the task coverage to complex multi-shot sequencing and audio-visual integration. Consequently, by providing granular diagnostic signals, EvalVerse transcends a static leaderboard and establishes a fundamental infrastructure for future work, such as reward models and evaluator agent.

cs.CV↗