arXiv subjects

Wenhao Sun

Publications and source records attributed to Wenhao Sun.

At least 19 recordsLinked to original sources

arXiv2026

PerturbMap: Cross-Context Transfer of Single-Cell Perturbation Responses

Single-cell perturbation atlases rarely measure every intervention in every cellular context: a query perturbation is often observed in one or more source contexts but missing in the recipient context where its effect is needed. Ignoring those measured responses discards query-specific experimental evidence, whereas copying or weakly calibrating them across contexts risks transferring the wrong signal. We propose PerturbMap, which predicts a missing recipient-context effect by combining a recipient-local low-rank base with accepted proposals that transport the same perturbation's measured source responses through source-to-recipient ridge experts fit on paired training perturbations, with proposal weights determined by route reliability estimated on validation anchors. On the Perturb-CITE-seq melanoma cohort, PerturbMap improves full-effect MSE by 4.1\% over a recipient-local low-rank base and achieves lower MSE than FedAvg, zero-response, raw-copy, calibrated-copy, and identity-shuffled affine controls. It remains within $2.82\times10^{-6}$ MSE of our centralized token-matched pooled reference, which uses a stronger training interface. A condition-mean specificity diagnostic shows the same direction: same-recipient top-10 counterpart retrieval by cosine increases from 74.5\% for the low-rank base to 80.5\% for PerturbMap.

cs.AI↗

arXiv2026

Origin of the reaction temperature in solid-state materials synthesis

Temperature plays a crucial role in solid-state materials synthesis, but there is currently no mechanistic theory to explain or predict which temperature is best to conduct a solid-state reaction. Reactions between powder precursors are conventionally assumed to be slow diffusion-limited processes; however, recent in situ experiments show that solid-state reactions can complete in minutes above a critical onset temperature. Here, we present evidence that a transient liquid phase forms above the metastable eutectic temperature, and that this non-equilibrium liquid serves as a fast diffusion medium to intermix precursors and initiate a solid-state reaction. This thermodynamic principle is agnostic to the structure or chemistry of the reactants, and can be applied towards the synthesis and manufacturing of a wide range of complex materials.

cond-mat.mtrl-sci↗

arXiv2026

Wavelet-Guided Semantic Signal Compensation for Inversion-Free Image Editing

Text-guided image editing aims to modify visual content according to a target prompt while preserving the background. Recent inversion-free image editing frameworks such as FlowEdit have demonstrated strong editing capability without requiring inversion. Empirically, FlowEdit can achieve substantial semantic changes under appropriate hyperparameter settings. However, we observe that under certain global attribute shifts, the editing trajectory may not effectively move away from the source distribution in the early timesteps. Our analysis suggests that in the high-noise regime, the dominant manifold-seeking flow toward the data manifold can reduce the influence of the text-conditioned direction, leading to limited global modification while background structures remain only moderately preserved. Inspired by this observation, we propose an inversion-free, frequency-aware semantic compensation strategy that strengthens the effective signal in the early stage of generation, while maintaining structural consistency in the background. The proposed method improves global editing capacity without sacrificing background fidelity.

cs.CV↗

arXiv2026

AutoPKG: An Automated Framework for Dynamic E-commerce Product-Attribute Knowledge Graph Construction

Product attribute extraction in e-commerce is bottlenecked by ontologies that are inconsistent, incomplete, and costly to maintain. We present AutoPKG, a multi-agent Large Language Model (LLM) framework that automatically constructs a Product-attribute Knowledge Graph (PKG) from multimodal product content. AutoPKG induces product types and type-specific attribute keys on demand, extracts attribute values from text and images, and consolidates updates through a centralized decision agent that maintains a globally consistent canonical graph. We also propose an evaluation protocol for dynamic PKGs that measures type and key validity, consolidation quality, and edge-level accuracy for value assertions after canonicalization. On a large real-world marketplace catalog dataset from Lazada (Alibaba), AutoPKG achieves up to 0.953 Weighted Knowledge Efficiency (WKE) for product types, 0.724 WKE for attribute keys, and 0.531 edge-level F1 for multimodal value extraction. Across three public benchmarks, our method improves edge-level exact-match F1 by 0.152 and yields a precision gain of 0.208 on the attribute extraction application. Online A/B tests show that AutoPKG-derived attributes increase Gross Merchandise Value (GMV) in Badge by 3.81 percent, in Search by 5.32 percent, and in Recommendation by 7.89 percent, supporting the practical value of AutoPKG in production.

cs.AI↗

arXiv2026

The Costs of Early-career Disciplinary Pivots: Evidence from Ph.D. Admissions

Scientific innovation often comes from researchers who pivot across disciplines. However, prior work found that established researchers face productivity penalties when pivoting. Here, we investigate the consequences of pivoting at the beginning of a research career -- doctoral admissions -- when the benefits of importing new ideas might outweigh the switching costs. Using applications to all PhD programs at a large research-intensive university between 2013-2023, we find that pivoters (those applying to programs outside their prior disciplinary training) have lower GPAs and standardized test scores than non-pivoters. Yet even conditional on these predictors of admission, pivoters are 1.3 percentage points less likely to be admitted. Examining applicants who applied to multiple programs in the same admissions cycle provides suggestive evidence that the admissions pivot penalty is causal. This penalty is significantly smaller for applicants who secure a recommendation from someone within the target discipline. Among those admitted and enrolled, pivoters are 12.9 percentage points less likely to graduate and do not show superior publication performance on average or at the tail. Our results reveal the substantial costs of disciplinary pivoting even at the outset of research careers, which constrain the flow of new ideas into research communities.

econ.GN↗

arXiv2026

Just-in-Time: Training-Free Spatial Acceleration for Diffusion Transformers

Diffusion Transformers have established a new state-of-the-art in image synthesis, but the high computational cost of iterative sampling severely hampers their practical deployment. While existing acceleration methods often focus on the temporal domain, they overlook the substantial spatial redundancy inherent in the generative process, where global structures emerge long before fine-grained details are formed. The uniform computational treatment of all spatial regions represents a critical inefficiency. In this paper, we introduce Just-in-Time (JiT), a novel training-free framework that addresses this challenge by acceleration in the spatial domain. JiT formulates a spatially approximated generative ordinary differential equation (ODE) that drives the full latent state evolution based on computations from a dynamically selected, sparse subset of anchor tokens. To ensure seamless transitions as new tokens are incorporated to expand the dimensions of the latent state, we propose a deterministic micro-flow, a simple and effective finite-time ODE that maintains both structural coherence and statistical correctness. Extensive experiments on the state-of-the-art FLUX.1-dev model demonstrate that JiT achieves up to a 7x speedup with nearly lossless performance, significantly outperforming existing acceleration methods and establishing a new and superior trade-off between inference speed and generation fidelity.

cs.CV↗

arXiv2026

MixerCSeg: An Efficient Mixer Architecture for Crack Segmentation via Decoupled Mamba Attention

Feature encoders play a key role in pixel-level crack segmentation by shaping the representation of fine textures and thin structures. Existing CNN-, Transformer-, and Mamba-based models each capture only part of the required spatial or structural information, leaving clear gaps in modeling complex crack patterns. To address this, we present MixerCSeg, a mixer architecture designed like a coordinated team of specialists, where CNN-like pathways focus on local textures, Transformer-style paths capture global dependencies, and Mamba-inspired flows model sequential context within a single encoder. At the core of MixerCSeg is the TransMixer, which explores Mamba's latent attention behavior while establishing dedicated pathways that naturally express both locality and global awareness. To further enhance structural fidelity, we introduce a spatial block processing strategy and a Direction-guided Edge Gated Convolution (DEGConv) that strengthens edge sensitivity under irregular crack geometries with minimal computational overhead. A Spatial Refinement Multi-Level Fusion (SRF) module is then employed to refine multi-scale details without increasing complexity. Extensive experiments on multiple crack segmentation benchmarks show that MixerCSeg achieves state-of-the-art performance with only 2.05 GFLOPs and 2.54 M parameters, demonstrating both efficiency and strong representational capability. The code is available at https://github.com/spiderforest/MixerCSeg.

cs.CV↗

arXiv2026

Thin-Film Stabilization and Magnetism of {\eta}-Carbide Type Iron Nitrides

Transition-metal nitrides in {\eta}-carbide type structures exhibit unusual bonding motifs and proximity to magnetic instabilities. Yet they remain unexplored in thin-film form due to the difficulty of stabilizing nitrogen-poor ternaries among competing phases. Here, we report the thin-film synthesis and phase-stability mapping of the {\eta}-nitride systems Fe-W-N and Fe-Mo-N. Amorphous Fe-M-N (M = W, Mo) combinatorial libraries deposited by reactive co-sputtering crystallize upon rapid thermal annealing, enabling systematic identification of synthesis windows as a function of composition and annealing temperature. Using laboratory powder X-ray diffraction and synchrotron grazing incidence wide angle X-ray scattering, we establish that Fe3Mo3N-based {\eta}-carbide phases form over a substantially broader compositional and thermal range than W-based compositions, where {\eta} structures are stabilized only when the films are Fe-rich. These trends are rationalized using mixed chemical-potential vs. composition phase diagrams that capture the narrow nitrogen chemical-potential stability of {\eta}-nitrides. Magnetic measurements reveal that ferromagnetism is induced in Fe-rich Fe3.54Mo2.46N with a small exchange-bias-like response that is absent in Fe3W3N-based compositions, highlighting the sensitivity of magnetic behavior to modest deviations from stoichiometry. This work establishes practical thin-film synthesis routes for {\eta}-nitride materials and demonstrates how composition can be tuned to access emergent magnetic phenomena in these complex nitrides.

cond-mat.mtrl-sci↗

arXiv2026

Ferroelectric dynamic-field-driven nucleation and growth model for predictive materials-to-circuit co-design

Real ferroelectric devices operate under mixed and distorted time-varying voltages, yet the standard nucleation-growth frameworks used to interpret ferroelectric switching - most notably the Kolmogorov-Avrami-Ishibashi (KAI) and nucleation-limited switching models (NLS) - are derived under the critically limiting assumption of a constant electric field. Thus, the prevailing interpretation of ferroelectric switching dynamics fails under real operating conditions. Here we introduce a compact dynamic-field-driven nucleation and growth (DFNG) model that enables quantitative fits to switching transients across multiple ferroelectric materials to extract time-varying domain wall velocity and growth dimensionality, even under arbitrary voltage waveform. This capability then motivates its use in device modeling under complex signals spanning disparate time and frequency scales. Coupling the compact model to application-related waveforms and circuit-level simulation platform facilitates a predictive materials-circuit co-design framework by linking nucleation and growth parameters to memory window, disturb error, speed, and energy dissipation for next-generation ferroelectric technologies.

cond-mat.mtrl-sci↗

arXiv2026

SPA-Cache: Singular Proxies for Adaptive Caching in Diffusion Language Models

While Diffusion Language Models (DLMs) offer a flexible, arbitrary-order alternative to the autoregressive paradigm, their non-causal nature precludes standard KV caching, forcing costly hidden state recomputation at every decoding step. Existing DLM caching approaches reduce this cost by selective hidden state updates; however, they are still limited by (i) costly token-wise update identification heuristics and (ii) rigid, uniform budget allocation that fails to account for heterogeneous hidden state dynamics. To address these challenges, we present SPA-Cache that jointly optimizes update identification and budget allocation in DLM cache. First, we derive a low-dimensional singular proxy that enables the identification of update-critical tokens in a low-dimensional subspace, substantially reducing the overhead of update identification. Second, we introduce an adaptive strategy that allocates fewer updates to stable layers without degrading generation quality. Together, these contributions significantly improve the efficiency of DLMs, yielding up to an $8\times$ throughput improvement over vanilla decoding and a $2$--$4\times$ speedup over existing caching baselines.

cs.LG↗

arXiv2025

Time-Temperature-Transformation (TTT) Diagrams to rationalize the nucleation and quenchability of metastable $\alpha$-Li$_3$PS$_4$

$\alpha$-Li$_3$PS$_4$ is a promising solid-state electrolyte with the highest ionic conductivity among its polymorphs. However, its formation presents a thermodynamic paradox: the $\alpha$-phase is the equilibrium phase at high temperature and transforms to the stable $\gamma$-Li$_3$PS$_4$ polymorph when cooled to room temperature; however, $\alpha$-Li$_3$PS$_4$ can be synthesized and quenched in a metastable state via rapid heating at relatively low temperatures. The origin of this synthesizability and anomalous stability has remained elusive. Here, we resolve this paradox by establishing a comprehensive time-temperature-transformation (TTT) diagram, constructed from a computational temperature-size phase diagram and experimental high-time-resolution isothermal measurements. Our density functional theory calculations reveal that at the nanoscale, the $\alpha$-phase is stabilized by its low surface energy, which drastically lowers the nucleation barrier across a wide temperature range. This size-dependent stabilization is directly visualized using in-situ synchrotron X-ray diffraction and electron microscopy, capturing the rapid nucleation of nano-sized $\alpha$-phase and its subsequent slow transformation. This work presents a generalizable framework that integrates thermodynamic and kinetic factors for understanding nucleation and phase transformation mechanisms, providing a rational strategy for the targeted synthesis of functional metastable materials.

cond-mat.mtrl-sci↗

arXiv2025

Magnetosynthesis effect on the structure and ground state of Cu$^{2+}$-based antiferromagnets

Subtle synthetic variables can have an outsizes influence on the crystal structure and magnetic properties of a material, particularly those of quantum materials. In this work, we investigate the impact of synthesis under a magnetic field (magnetosynthesis) on the crystal structure and magnetic properties of several Cu$^{2+}$ ($S=1/2$) based materials with antiferromagnetic interactions and varying levels of magnetic frustration, from simple antiferromagnets to a quantum spin liquid. We employ small (0.09 - 0.37 T) magnetic fields applied during low-temperature hydrothermal or evaporative synthesis of the simple antiferromagnet CuCl$_2\cdot$2H$_2$O, the canted antiferromagnet (Cu,Zn)$_3$Cl$_4$(OH)$_2\cdot$2H$_2$O, the frustrated and canted antiferromagnet atacamite Cu$_2$(OH)$_3$Cl, and the highly frustrated quantum spin liquid herbertsmithite Cu$_3$Zn(OH)$_6$Cl$_2$. We found that (Cu,Zn)$_3$Cl$_4$(OH)$_2\cdot$2H$_2$O experiences structural changes well above its magnetic transition. Atacamite Cu$_2$(OH)$_3$Cl synthesized under a 0.19 T field experiences a 0.15 K (~3%) decrease in its N\'eel transition temperature and a significant strengthening of its antiferromagnetic interactions, suggesting that magnetosynthesis can influence the ground state of moderately frustrated materials.

cond-mat.str-el↗

arXiv2025

Crystal Orbital Guided Iteration to Atomic Orbitals: A Pathway to Chemically Adaptive Atomic Orbitals from DFT

Atomic orbitals underpin our understanding of electronic structure, providing intuitive descriptions of bonding, charge transfer, magnetism, and correlation effects. Despite their utility, an atomic basis that is adaptable, strictly localized on atomic centers, and enables accurate tight-binding interpolation has remained elusive. Here, we introduce Crystal Orbital Guided Iteration To atomic-Orbitals (COGITO), a framework that constructs an optimal atomic orbital basis by identifying and resolving key mathematical obstacles inherent to nonorthogonal bases -- particularly uncontrolled orbital mixing, and the fixed-overlap constraint between orbitals. We demonstrate that COGITO enables tight-binding models as accurate as MLWF-based approaches, while preserving the ability of tight-binding parameters to represent the projected atomic basis -- an essential feature lost in schemes that enforce orbital orthogonality or maximal localization. By creating accurate and chemically interpretable models of electronic structure, COGITO reveals the orbital-resolved covalent bonds and charge transfer that is encoded in the Kohn-Sham wavefunctions of DFT. Our method thus offers a powerful tool for any physics- or chemistry-based application that relies on a faithful description of local electronic structure.

cond-mat.mtrl-sci↗

arXiv2025

Lightning Fast Caching-based Parallel Denoising Prediction for Accelerating Talking Head Generation

Diffusion-based talking head models generate high-quality, photorealistic videos but suffer from slow inference, limiting practical applications. Existing acceleration methods for general diffusion models fail to exploit the temporal and spatial redundancies unique to talking head generation. In this paper, we propose a task-specific framework addressing these inefficiencies through two key innovations. First, we introduce Lightning-fast Caching-based Parallel denoising prediction (LightningCP), caching static features to bypass most model layers in inference time. We also enable parallel prediction using cached features and estimated noisy latents as inputs, efficiently bypassing sequential sampling. Second, we propose Decoupled Foreground Attention (DFA) to further accelerate attention computations, exploiting the spatial decoupling in talking head videos to restrict attention to dynamic foreground regions. Additionally, we remove reference features in certain layers to bring extra speedup. Extensive experiments demonstrate that our framework significantly improves inference speed while preserving video quality.

cs.GR↗

arXiv2025

AD-FM: Multimodal LLMs for Anomaly Detection via Multi-Stage Reasoning and Fine-Grained Reward Optimization

While Multimodal Large Language Models (MLLMs) demonstrate remarkable capabilities across diverse domains, their application to specialized anomaly detection (AD) remains constrained by domain adaptation challenges. Existing Group Relative Policy Optimization (GRPO) based approaches suffer from two critical limitations: inadequate training data utilization when models produce uniform responses, and insufficient supervision over reasoning processes that encourage immediate binary decisions without deliberative analysis. We propose a comprehensive framework addressing these limitations through two synergistic innovations. First, we introduce a multi-stage deliberative reasoning process that guides models from region identification to focused examination, generating diverse response patterns essential for GRPO optimization while enabling structured supervision over analytical workflows. Second, we develop a fine-grained reward mechanism incorporating classification accuracy and localization supervision, transforming binary feedback into continuous signals that distinguish genuine analytical insight from spurious correctness. Comprehensive evaluation across multiple industrial datasets demonstrates substantial performance improvements in adapting general vision-language models to specialized anomaly detection. Our method achieves superior accuracy with efficient adaptation of existing annotations, effectively bridging the gap between general-purpose MLLM capabilities and the fine-grained visual discrimination required for detecting subtle manufacturing defects and structural irregularities.

cs.CV↗

arXiv2025

SPAZER: Spatial-Semantic Progressive Reasoning Agent for Zero-shot 3D Visual Grounding

3D Visual Grounding (3DVG) aims to localize target objects within a 3D scene based on natural language queries. To alleviate the reliance on costly 3D training data, recent studies have explored zero-shot 3DVG by leveraging the extensive knowledge and powerful reasoning capabilities of pre-trained LLMs and VLMs. However, existing paradigms tend to emphasize either spatial (3D-based) or semantic (2D-based) understanding, limiting their effectiveness in complex real-world applications. In this work, we introduce SPAZER - a VLM-driven agent that combines both modalities in a progressive reasoning framework. It first holistically analyzes the scene and produces a 3D rendering from the optimal viewpoint. Based on this, anchor-guided candidate screening is conducted to perform a coarse-level localization of potential objects. Furthermore, leveraging retrieved relevant 2D camera images, 3D-2D joint decision-making is efficiently performed to determine the best-matching object. By bridging spatial and semantic reasoning neural streams, SPAZER achieves robust zero-shot grounding without training on 3D-labeled data. Extensive experiments on ScanRefer and Nr3D benchmarks demonstrate that SPAZER significantly outperforms previous state-of-the-art zero-shot methods, achieving notable gains of 9.0% and 10.9% in accuracy.

cs.CV↗

arXiv2025

Heavier chalcogenofenchones for fundamental gas-phase studies of molecular chirality

Monoterpene ketones are frequently studied compounds that enjoy great popularity both in chemistry and in physics due to comparatively high volatility, stability, conformational rigidity and commercial availability. Herein, we explore the heavier chalcogenoketone derivatives of fenchone as promising benchmark systems -- synthetically accessible in enantiomerically pure form -- for systematic studies of nuclear charge ($Z$) dependent properties in chiral compounds. Synthesis, structural characterization, thorough gas-phase rotational and vibrational spectroscopy as well as accompanying quantum chemical studies on the density-functional-theory level reported in this work foreshadow subsequent applications of this compound class for fundamental investigations of molecular chirality under well-defined conditions.

physics.chem-ph↗

arXiv2025

Multimodal Reasoning Agent for Zero-Shot Composed Image Retrieval

Zero-Shot Composed Image Retrieval (ZS-CIR) aims to retrieve target images given a compositional query, consisting of a reference image and a modifying text-without relying on annotated training data. Existing approaches often generate a synthetic target text using large language models (LLMs) to serve as an intermediate anchor between the compositional query and the target image. Models are then trained to align the compositional query with the generated text, and separately align images with their corresponding texts using contrastive learning. However, this reliance on intermediate text introduces error propagation, as inaccuracies in query-to-text and text-to-image mappings accumulate, ultimately degrading retrieval performance. To address these problems, we propose a novel framework by employing a Multimodal Reasoning Agent (MRA) for ZS-CIR. MRA eliminates the dependence on textual intermediaries by directly constructing triplets, , using only unlabeled image data. By training on these synthetic triplets, our model learns to capture the relationships between compositional queries and candidate images directly. Extensive experiments on three standard CIR benchmarks demonstrate the effectiveness of our approach. On the FashionIQ dataset, our method improves Average R@10 by at least 7.5\% over existing baselines; on CIRR, it boosts R@1 by 9.6\%; and on CIRCO, it increases mAP@5 by 9.5\%.

cs.CV↗