Searcharxiv⌕ Search

arXiv subjects

Feng Zhang

Publications and source records attributed to Feng Zhang.

At least 37 records · Page 2Linked to original sources
arXiv2026

DiffImaginE: Imagine to Verify Entity Types with Diffusion

Multimodal named entity recognition (MNER) determines whether each candidate span and entity-type hypothesis is supported by joint textual and visual evidence. Existing imagine-and-compare verifiers map each (span, type) pair to one predicted visual feature, compressing diverse visual realisations into a single prototype and providing a compatibility score without explicit probabilistic semantics. We introduce DiffImaginE, which formulates MNER type verification as conditional latent diffusion inference. Given span-localised visual evidence, a type-conditioned denoiser predicts noise injected into its standardised latent. The resulting denoising error provides an ELBO-consistent surrogate for type-conditional negative log-likelihood, allowing competing type hypotheses to be ranked by how well they explain the observation. DiffImaginE retains a standard multimodal encoder stack and replaces the deterministic verifier with a classifier-free-guided diffusion scorer trained using Min-SNR weighting. We directly supervise per-type diffusion scores as classification logits, learn aggregation across noise levels, and use antithetic sampling to reduce Monte Carlo comparison variance. Our analysis shows that classifier-free guidance sharpens the induced type posterior and characterises when antithetic pairing reduces variance at equal denoiser cost. Experiments on Twitter-2015 and Twitter-2017 show consistent gains over a matched deterministic ImaginE control under the same encoder, auxiliary objectives, and evaluation protocol, supported by ablations and paired significance tests.

cs.AI↗
arXiv2026

DocPO: Advancing Document Policy Optimization via Tailored Step-Aware Rewards

Reinforcement learning (RL) for document parsing often relies on reference-based rewards rooted in edit distance (e.g., tree edit distance), yet it remains hard to optimize in the high-accuracy regime because such rewards become weakly discriminative: near-correct outputs receive very similar scores, providing limited learning signal for hard cases. We propose Step-Aware Annealing (SAA), a plug-and-play reward sharpening mechanism that progressively increases reward curvature during training, amplifying subtle quality differences among high-scoring samples while preserving stability in early learning. Built on SAA, we introduce DocPO, a document policy optimization framework with element-specific, reference-based rewards anchored by edit-distance signals: normalized string edit distance (NED) for text, tree edit distance similarity (TEDS) for tables, and a hybrid Rubric+edit reward for formulas. Experiments on OmniDocBench and DocElemHard show that SAA consistently improves GRPO-style RL across document elements over non-annealed rewards, without requiring additional human supervision for reward construction.

cs.CV↗
arXiv2026

Scalable Frequency- and Length-Aware Subdocument Deduplication for Large Language Model Pretraining

Large-scale pretraining corpora contain substantial duplicate content. Although document-level deduplication is widely used, removing subdocument-level redundancy remains challenging. At corpus scale, suffix-array-based methods are commonly applied independently within shards, leaving cross-shard duplicates undetected and making the resulting retention behavior sensitive to the sharding configuration. Hash-based methods enable global exact duplicate counting, but often rely on fixed copy-retention policies that cannot accommodate heterogeneous repetition patterns. We propose a scalable subdocument deduplication framework that decouples duplicate detection from copy retention. It identifies duplicate groups through natural-boundary segmentation, normalized exact hashing, and distributed aggregation, and then applies an explicit frequency- and length-aware retention policy that allocates an adaptive copy budget to each group, retaining more copies of low-frequency or short repetitions while more aggressively deleting high-frequency or long ones. Experiments on FineWeb-Edu and a code-containing web corpus show that models trained on data processed by our method achieve the best overall performance among the evaluated settings. These results underscore the importance of explicit copy-retention control.

cs.CL↗
arXiv2026

Projection-shifted particle-flow imaging with cosmic-ray muons

Cosmic-ray muons are natural probes for non-destructive imaging, but reaching sub-millimetre resolution with realistic exposure times has long been hindered by a fundamental limitation: the stochastic nature of multiple Coulomb scattering defies deterministic reconstruction of particle trajectories inside matter. Here we introduce Projection-shifted MUon transMission tomogrAphy (P$μ$MA). Rather than localizing individual scattering points, P$μ$MA records how material perturbations statistically shift the projected transmission tracks of a muon flux onto a virtual imaging plane. Near density boundaries, asymmetric projection-shift statistics naturally produce steep undershoot-overshoot profiles, sharpening edges without additional filtering. Crucially, the method operates robustly even with only two detector planes, a configuration where conventional scattering tomography cannot be applied. Cosmic-ray simulations with a lead knife-edge target yield edge widths as narrow as 1.196 mm, while monoenergetic-beam simulations reach 48 $μ$m. Using a prototype system, we resolve 2-mm copper letters within two days---a feat unattainable by standard approaches under equivalent conditions. The projection-shift concept can be extended to accelerator- or laser-driven muon beams and to other ions, establishing a generalisable strategy for high-resolution particle-flow imaging.

physics.ins-det↗
arXiv2026

Deep Research Pretraining via Predictive Navigation

Deep research agents are often trained on expensive, environment-grounded tool-use trajectories that require repeated retrieval, document inspection, and report evaluation. We introduce Deep Research Pretraining (DRP), an offline framework that derives predictive navigation supervision from naturally occurring evidence structures. Given a citation-bearing or hyperlinked passage, DRP constructs a proxy research objective, recovers linked evidence and graph-related alternatives, and converts them into search-open-write trajectories. This teaches models what to search for, which documents to inspect, and how to synthesize evidence, without a live retrieval environment or executed policy rollout. We instantiate DRP on scholarly citation graphs (DRP-Paper) and Wikipedia hyperlinks (DRP-Web), continually pretrain separate Qwen3-14B-Base models on 1B tokens, and fine-tune them on controlled fractions of 13K agent trajectories. Across five independently sampled subsets at each low-data budget, both variants consistently outperform matched no-DRP models on DeepResearch Bench. With one quarter of the SFT data, DRP-Web even surpasses a fixed no-DRP full-data checkpoint, with gains transferring to ResearchQA, WebWalkerQA, and SimpleQA. Starting from matched low-data SFT checkpoints, the DRP-Web advantage also persists through subsequent agentic RL. Source-matched and evidence-mismatch controls indicate that these improvements arise from evidence-conditioned navigation rather than domain exposure or agent-format imitation. DRP thus provides a promising complementary approach to trajectory-based agent training.

cs.CL↗
arXiv2026

One4Many-StablePacker: An Efficient Deep Reinforcement Learning Framework for the 3D Bin Packing Problem

The three-dimensional bin packing problem (3D-BPP) is widely applied in logistics and warehousing. Existing learning-based approaches often neglect practical stability-related constraints and exhibit limitations in generalizing across diverse bin dimensions. To address these limitations, we propose a novel deep reinforcement learning framework, One4Many-StablePacker (O4M-SP). The primary advantage of O4M-SP is its ability to handle various bin dimensions in a single training process while incorporating support and weight constraints common in practice. Our training method introduces two innovative mechanisms. First, it employs a weighted reward function that integrates loading rate and a new height difference metric for packing layouts, promoting improved bin utilization through flatter packing configurations. Second, it combines clipped policy gradient optimization with a tailored policy drifting method to mitigate policy entropy collapse, encouraging exploration at critical decision nodes during packing to avoid suboptimal solutions. Extensive experiments demonstrate that O4M-SP generalizes successfully across diverse bin dimensions and significantly outperforms baseline methods. Furthermore, O4M-SP exhibits strong practical applicability by effectively addressing packing scenarios with stability constraints.

cs.LG↗
arXiv2026

The possible $K^{*}Σ^{*}$ molecular state

Within the framework of the one-boson-exchange model, we systematically investigate the interaction between the vector meson $K^{*}$ and the baryon $Σ^{*}$ with the aim of exploring the possibility of forming hadronic molecular states. The $K^{*}Σ^{*}$ interaction potential is constructed from $ρ$, $ω$, and $π$ meson exchanges, and the nonrelativistic Schrödinger equation is solved using the Gaussian expansion method. The binding energies are calculated for different total angular momenta $J^{P}$ and isospin channels $I=1/2$ and $I=3/2$. Our results show that $S$--$D$ wave mixed $K^{*}Σ^{*}$ molecular states with $J^{P}=1/2^{-}$ can be formed only in the $I=3/2$ channel, while no bound state appears in the $I=1/2$ channel. In addition, the $S$--$D$ wave mixed states with $J^{P}=3/2^{-}$ and $J^{P}=5/2^{-}$ are also found to support bound-state solutions. For higher partial-wave states in our study, the binding mechanism mainly arises from the interplay between partial-wave mixing and non-central interactions. In particular, the $J^{P}=1/2^{+}$ channel does not support a bound state, as the meson-exchange interaction is predominantly repulsive. Our analysis further supports the interpretation of the experimentally observed $N(2250)$ and $Δ(2200)$ states as $K^{*}Σ^{*}$ molecular candidates, corresponding to $I=1/2,\ J^{P}=9/2^{-}$ and $I=3/2,\ J^{P}=7/2^{-}$, respectively.

hep-ph↗
arXiv2026

Alchemical thermodynamic integration for ab initio free-energy calculations in solutions

We develop an alchemical thermodynamic integration scheme that couples ab initio force calculators on the fly during Monte Carlo and molecular dynamics simulations. The implementation is validated against existing hybrid-Hamiltonian approaches. The scheme yields ab initio free energies of high-pressure Fe-Ni and ambient Li-Na liquid solutions that agree with previous calculations and reproduce the experimentally observed Li-Na miscibility gap. The code has an efficiency comparable to standard ab initio molecular dynamics. These results establish this scheme as a practical alchemical-integration framework for first-principles free-energy calculations in solutions.

cond-mat.mtrl-sci↗
arXiv2026

exa-PD: A scalable high-performance workflow for multi-element phase diagram construction

Exa-PD is a highly parallelizable workflow designed for the construction of multi-element phase diagrams (PDs). It uses standard sampling techniques, molecular dynamics (MD) and Monte Carlo (MC) as implemented in the LAMMPS package, to simultaneously sample multiple phases over a fine temperature-composition mesh for free-energy calculations. Parsl serves as the global workflow engine, coordinating large ensembles of MD and MC tasks to achieve massive parallelization with strong scalability. The resulting free energies of liquid and solid phases are then fed to CALPHAD modeling via the PyCalphad package to construct multi-element PDs.

cond-mat.mtrl-sci↗
arXiv2026

VSC: A Zero-Dimensional Fusion Design Platform for Multiple Magnetic Configurations

The VeloAlpha System Code (VSC) is a computational framework for zero-dimensional fusion power-balance studies across five magnetic-confinement configurations: tokamaks, magnetic mirrors, field-reversed configurations (FRCs), dipoles, and stellarators. A common power-balance formulation connects fusion production, charged-particle deposition, radiation, transport loss, external heating, and fusion gain, while each configuration retains its own geometry, profile weights, confinement model, and operating constraints. The same solver interface supports both single-point calculations and two-dimensional plasma operating contour (POPCON) scans, producing fusion and heating powers, gain, radiation and transport losses, geometry quantities, and configuration-specific validity indicators. VSC therefore makes it possible to study how assumptions about density, temperature, magnetic field, confinement, and geometry shape the accessible operating space of different fusion concepts within one traceable framework. By combining reduced-order physics models with a unified computational platform, VSC enables rapid assessment and comparative analysis of candidate fusion reactor concepts during the early design stage.

physics.plasm-ph↗
arXiv2026

DeltaDeno: Zero-Shot Anomaly Generation via Delta-Denoising Attribution

Anomaly generation is often framed as few-shot fine-tuning with anomalous samples, which contradicts the scarcity that motivates generation and tends to overfit category priors. We tackle the setting where no real anomaly samples or training are available. We propose Delta-Denoising (\textbf{DeltaDeno}), a training-free zero-shot anomaly generation method that localizes and edits defects by contrasting two diffusion branches driven by a minimal prompt pair under a shared schedule. By accumulating per-step denoising deltas into an image-specific localization map, we obtain a mask to guide the latent inpainting during later diffusion steps and preserve the surrounding context while generating realistic local defects. To improve stability and control, DeltaDeno performs token-level prompt refinement that aligns shared content and strengthens anomaly tokens, and applies a spatial attention bias restricted to anomaly tokens in the predicted region. Experiments on public datasets show that DeltaDeno achieves great generation, realism and consistent gains in downstream detection performance. Code will be made publicly available at https://github.com/CROVO1026/DeltaDeno.

cs.CV↗
arXiv2026

CARL: Constraint-Aware Reinforcement Learning for Planning with LLMs

Despite their strong reasoning capabilities and extensive world knowledge, Large Language Models (LLMs) frequently generate plans that violate task constraints, undermining their reliability in real-world applications. This deficiency arises from a lack of systematic mechanisms to incorporate constraint information during the generation process. While existing approaches attempt to mitigate this by relying on external tools or task decomposition, they fail to enhance the model's intrinsic constraint awareness. To address this, we propose Constraint-Aware Reinforcement Learning (CARL), a novel RL framework designed to strengthen LLMs' intrinsic focus on constraints. CARL introduces a constraint-aware reward by comparing the model's output distributions under constrained and unconstrained inputs, encouraging constraint focus and penalizing neglect. Compatible with various RL frameworks and requiring no external solvers or top models, CARL enables scalable, end-to-end constraint-aware planning. Extensive experiments on BlocksWorld, TravelPlanner, and T-Eval demonstrate that CARL significantly outperforms standard Reinforcement Fine-Tuning (RFT) baselines and state-of-the-art reasoning models, exhibiting a markedly increased focus on constraints.

cs.AI↗
arXiv2026

STAPO: Selective Trajectory-Aware Policy Optimization for LLM Agent Training

Reinforcement Learning (RL) is the dominant paradigm for training Large Language Model (LLM) agents on long-horizon tasks. However, sparse and delayed rewards often lead to trajectory neglect, in which agents lose focus on the task goal and interaction history at intermediate steps. Prior work has explored step-level supervision using Shannon-entropy-based uncertainty signals, which conflate inherent state complexity with agent confidence and therefore provide unreliable estimates of decision reliability. To address this issue, we propose normalized entropy, which measures confidence deviations relative to an agent's average behavior under a given state, thereby strengthening the association between low-quality actions and trajectory neglect. Building on this insight, we introduce Selective Trajectory-Aware Policy Optimization (STAPO), a hierarchical group-based RL framework. STAPO leverages normalized entropy to locate outlier steps associated with trajectory neglect and optimizes them via a joint mechanism of trajectory-aware reward and trajectory-independent penalty, enhancing trajectory awareness while preserving training stability. Extensive experiments on ALFWorld, WebShop, and Search-Augmented QA demonstrate that STAPO achieves state-of-the-art performance while substantially alleviating trajectory neglect, validating its effectiveness and robustness for agentic tasks.

cs.AI↗
arXiv2026

MORE: A Multilingual Document Parsing Benchmark and Evaluation

Multilingual documents encapsulate rich regional cultures, scientific discoveries, and historical records. Parsing this content into structured, machine-readable formats is critical for unlocking global knowledge. However, existing benchmarks predominantly focus on high-resource languages like English and Chinese, creating an evaluation blind spot concerning model performance on other languages. While recent Vision-Language Models (VLMs) claim support for hundreds of languages, the lack of ground truth makes it impossible to empirically verify these capabilities. To bridge this gap, we introduce MORE, a large-scale benchmark designed for multilingual document parsing evaluation. MORE distinguishes itself through three key dimensions: (1) Unprecedented Scale: It covers 149 languages, making it the most linguistically diverse benchmark to date; (2) Structural Complexity: Unlike previous works, it extends evaluation beyond plain text to include structural elements such as code blocks, tables, and catalogs; and (3) Data Authenticity: All samples are curated from real-world documents via a model-assisted, human-refined annotation pipeline. We evaluate state-of-the-art models using MORE, establishing new performance baselines for long-tail languages and validating the benchmark's effectiveness in diagnosing model capabilities in realistic, diverse scenarios. The MORE dataset will be available at https://github.com/zimoqingfeng/MORE.

cs.CV↗
arXiv2026

Home3D 1.0: A High-Fidelity Image-to-3D Asset Generation System for Interior Design

We present Home3D 1.0, a modular image-to-3D generation system that produces high-quality 3D assets from a single reference image, targeting interior design and e-commerce applications. Given a photograph of a furniture or decor item, the system outputs a mesh with physically-based rendering (PBR) materials, and the mesh can be decomposed into material-specific components. The pipeline is organized into four tightly coupled modules: Geometry reconstructs a watertight mesh through latent SDF modelling with a geometry VAE and a coarse-to-fine flow-matching DiT; Texture predicts multiview albedo observations, reprojects them onto the mesh, and completes unseen surface regions with a 3D texture field; Material uses MatWeaver to obtain component masks through video-based segmentation and UV-space voting, then retrieves and bakes PBR maps from a curated material library through hierarchical multi-modal matching; and Parts generates material-editable semantic part meshes with a PartVAE and PartDiT, decoding multi-head part-specific SDF fields in one pass. Each module is evaluated independently with dedicated metrics, highlighting both the current system capability and the remaining gaps toward broader deployment.

cs.CV↗
arXiv2026

Low-Rank Tensor Completion Based on Fractional Regularization with Ky Fan p-k Norm

This paper addresses low-rank tensor completion (LRTC) by proposing a novel nonconvex surrogate, namely the ratio of the tensor nuclear norm to the tensor Ky Fan p-k norm (TNPK), to accurately approximate the tensor tubal rank. The TNPK possesses appealing properties, including scale invariance, parameter flexibility, and the existence of closed-form solutions under specific choices of p and k. With specific parameter settings of p and k, it reduces to the ratio of the tensor nuclear norm to the tensor Ky Fan k norm (TNK) or the ratio of the tensor nuclear norm to the tensor Frobenius norm (TNF). We construct a LRTC model and, under the tensor null space property (NSP), prove that low-rank tensors are local minimizers of the proposed model. Moreover, we derive the proximal operator of the Ky Fan p-k inverse-norm and further develop an efficient alternating direction method of multipliers (ADMM) algorithm with guaranteed subsequential convergence under mild conditions. Extensive experiments on synthetic and real-world datasets validate the superior performance of our method against state-of-the-art competitors.

cs.CV↗
arXiv2026

Bridging Passive and Active: Enhancing Conversation Starter Recommendation via Active Expression Modeling

Large Language Model (LLM)-driven conversational search is shifting information retrieval from reactive keyword matching to proactive, open-ended dialogues. In this context, Conversation Starters are widely deployed to provide personalized query recommendations that help users initiate dialogues. Conventionally, recommending these starters relies on a closed "exposure-click" loop. Yet, this feedback loop mechanism traps the system in an echo chamber where, compounded by data sparsity, it fails to capture the dynamic nature of conversational search intents shaped by the open world. As a result, the system skews towards popular but generic suggestions. In this work, we uncover an untapped paradigm shift to shatter this harmful feedback loop: harnessing user "free will" through active user expressions. Unlike traditional recommendations, conversational search empowers users to bypass menus entirely through manually typed queries. The open-world intents in active queries hold the key to breaking this loop. However, incorporating them is non-trivial: (1) there exists an inherent distribution shift between active queries and formulated starters. (2) Furthermore, the "non-ID-able" nature of open text renders traditional item-based popularity statistics ineffective for large-scale industrial streaming training. To this end, we propose Passive-Active Bridge (PA-Bridge), a novel framework that employs an adversarial distribution aligner to bridge the distributional gap between passively recommended starters and active expressions. Moreover, we introduce a semantic discretizer to enable the deployment of popularity debiasing algorithms. Online A/B tests on our platform, demonstrate that PA-Bridge significantly boosts the Feature Penetration Rate by 0.54% and User Active Days by 0.04%.

cs.IR↗
arXiv2026

EvoCut: Multi-Layer Evolution-Aware Visual Token Compression for Efficient Large Vision-Language Models

Large vision-language models (LVLMs) achieve strong performance on image and video understanding tasks, but their inference efficiency is constrained by the large number of visual tokens produced by vision encoders. Most existing visual token compression methods estimate token importance from attention scores or representation properties at specific layers, overlooking how visual tokens evolve across the vision encoder. Such layer-specific criteria may provide incomplete importance estimates and limit performance preservation after compression. To address this issue, we analyze layer-wise visual token evolution directions and observe that tokens form multiple group evolution directions across vision-encoder layers. Our analysis further shows that informative tokens tend to exhibit persistent deviations from common group evolution directions. Based on this observation, we propose EvoCut, a training-free and attention-free visual token compression method that estimates token importance from multi-layer evolution deviation. Experimental results show that EvoCut can retain only 11.1\% of the visual tokens on LLaVA-1.5-7B while preserving 94.4\% of the average performance, demonstrating its effectiveness in balancing efficiency and accuracy.

cs.CV↗