arXiv subjects

Lei Wang

Publications and source records attributed to Lei Wang.

At least 235 records · Page 13Linked to original sources

arXiv2026

Ab initio simulation of the first-order proton-ordering transition in water ice

Proton ordering in water ice is a paradigmatic order-disorder transition in a locally constrained system. The ice rules require exactly two hydrogens close to each oxygen, restricting the disorder to an exponentially large yet strongly correlated manifold of hydrogen-bond configurations. Within this constrained space, meV-scale energy differences drive the transition from disordered ice Ih to ordered ice XI, while distinct configurations are separated by eV-scale barriers. These barriers hinder equilibration in experiments, and efficient sampling of this space with the required energy accuracy has remained a long-standing challenge in simulation. We address this by combining a machine learning interatomic potential with loop updates that preserve the ice rules and continuous updates of atomic coordinates, enabling equilibrium sampling with ab initio accuracy and capturing configurational entropic effects. In systems of up to 360 water molecules, with over 10^6 samples retained per temperature point, the simulations reveal clear first-order transition signatures at 83 K: a negative Binder cumulant, a bimodal potential energy distribution, and a sharp step in the lattice aspect ratio. Nuclear quantum effects are estimated to lower the transition temperature by approximately 20 K, bringing the prediction closer to the experimental value of 72 K.

cond-mat.mtrl-sci↗

arXiv2026

WaDi: Weight Direction-aware Distillation for One-step Image Synthesis

Despite the impressive performance of diffusion models such as Stable Diffusion (SD) in image generation, their slow inference limits practical deployment. Recent works accelerate inference by distilling multi-step diffusion into one-step generators. To better understand the distillation mechanism, we analyze U-Net/DiT weight changes between one-step students and their multi-step teacher counterparts. Our analysis reveals that changes in weight direction significantly exceed those in weight norm, highlighting it as the key factor during distillation. Motivated by this insight, we propose the Low-rank Rotation of weight Direction (LoRaD), a parameter-efficient adapter tailored to one-step diffusion distillation. LoRaD is designed to model these structured directional changes using learnable low-rank rotation matrices. We further integrate LoRaD into Variational Score Distillation (VSD), resulting in Weight Direction-aware Distillation (WaDi)-a novel one-step distillation framework. WaDi achieves state-of-the-art FID scores on COCO 2014 and COCO 2017 while using only approximately 10% of the trainable parameters of the U-Net/DiT. Furthermore, the distilled one-step model demonstrates strong versatility and scalability, generalizing well to various downstream tasks such as controllable generation, relation inversion, and high-resolution synthesis.

cs.CV↗

arXiv2026

CCSD: Cross-Modal Compositional Self-Distillation for Robust Brain Tumor Segmentation with Missing Modalities

The accurate segmentation of brain tumors from multi-modal MRI is critical for clinical diagnosis and treatment planning. While integrating complementary information from various MRI sequences is a common practice, the frequent absence of one or more modalities in real-world clinical settings poses a significant challenge, severely compromising the performance and generalizability of deep learning-based segmentation models. To address this challenge, we propose a novel Cross-Modal Compositional Self-Distillation (CCSD) framework that can flexibly handle arbitrary combinations of input modalities. CCSD adopts a shared-specific encoder-decoder architecture and incorporates two self-distillation strategies: (i) a hierarchical modality self-distillation mechanism that transfers knowledge across modality hierarchies to reduce semantic discrepancies, and (ii) a progressive modality combination distillation approach that enhances robustness to missing modalities by simulating gradual modality dropout during training. Extensive experiments on public brain tumor segmentation benchmarks demonstrate that CCSD achieves state-of-the-art performance across various missing-modality scenarios, with strong generalization and stability.

cs.CV↗

arXiv2026

MCP-SafetyBench: A Benchmark for Safety Evaluation of Large Language Models with Real-World MCP Servers

Large language models (LLMs) are evolving into agentic systems that reason, plan, and operate external tools. The Model Context Protocol (MCP) is a key enabler of this transition, offering a standardized interface for connecting LLMs with heterogeneous tools and services. Yet MCP's openness and multi-server workflows introduce new safety risks that existing benchmarks fail to capture, as they focus on isolated attacks or lack real-world coverage. We present MCP-SafetyBench, a comprehensive benchmark built on real MCP servers that supports realistic multi-turn evaluation across five domains: browser automation, financial analysis, location navigation, repository management, and web search. It incorporates a unified taxonomy of 20 MCP attack types spanning server, host, and user sides, and includes tasks requiring multi-step reasoning and cross-server coordination under uncertainty. Using MCP-SafetyBench, we systematically evaluate leading open- and closed-source LLMs, revealing that all models remain vulnerable to MCP attacks, with a notable safety-utility trade-off. Our results highlight the urgent need for stronger defenses and establish MCP-SafetyBench as a foundation for diagnosing and mitigating safety risks in real-world MCP deployments.

cs.CL↗

arXiv2026

When Relevance Meets Novelty: Dual-Stable Periodic Optimization for Serendipitous Recommendation

Traditional recommendation systems tend to trap users in strong feedback loops by excessively pushing content aligned with their historical preferences, thereby limiting exploration opportunities and causing content fatigue. Although large language models (LLMs) demonstrate potential with their diverse content generation capabilities, existing LLM-enhanced dual-model frameworks face two major limitations: first, they overlook long-term preferences driven by group identity, leading to biased interest modeling; second, they suffer from static optimization flaws, as a one-time alignment process fails to leverage incremental user data for closed-loop optimization. To address these challenges, we propose the Co-Evolutionary Alignment (CoEA) method. For interest modeling bias, we introduce Dual-Stable Interest Exploration (DSIE) module, jointly modeling long-term group identity and short-term individual interests through parallel processing of behavioral sequences. For static optimization limitations, we design a Periodic Collaborative Optimization (PCO) mechanism. This mechanism regularly conducts preference verification on incremental data using the Relevance LLM, then guides the Novelty LLM to perform fine-tuning based on the verification results, and subsequently feeds back the output of the continually fine-tuned Novelty LLM to the Relevance LLM for re-evaluation, thereby achieving a dynamic closed-loop optimization. Extensive online and offline experiments verify the effectiveness of the CoEA model in serendipitous recommendation.

cs.IR↗

arXiv2026

Document Reconstruction Unlocks Scalable Long-Context RLVR

Reinforcement Learning with Verifiable Rewards~(RLVR) has become a prominent paradigm to enhance the capabilities (i.e.\ long-context) of Large Language Models~(LLMs). However, it often relies on gold-standard answers or explicit evaluation rubrics provided by powerful teacher models or human experts, which are costly and time-consuming. In this work, we investigate unsupervised approaches to enhance the long-context capabilities of LLMs, eliminating the need for heavy human annotations or teacher models' supervision. Specifically, we first replace a few paragraphs with special placeholders in a long document. LLMs are trained through reinforcement learning to reconstruct the document by correctly identifying and sequencing missing paragraphs from a set of candidate options. This training paradigm enables the model to capture global narrative coherence, significantly boosting long-context performance. We validate the effectiveness of our method on two widely used benchmarks, RULER and LongBench~v2. While acquiring noticeable gains on RULER, it can also achieve a reasonable improvement on LongBench~v2 without any manually curated long-context QA data. Furthermore, we conduct extensive ablation studies to analyze the impact of reward design, data curation strategies, training schemes, and data scaling effects on model performance. We publicly release our code, data, and models.

cs.CL↗

arXiv2026

Suppressing Prior-Comparison Hallucinations in Radiology Report Generation via Semantically Decoupled Latent Steering

Automated radiology report generation using vision-language models (VLMs) is limited by the risk of prior-comparison hallucination, where the model generates historical findings unsupported by the current study. We address this challenge with a training-free, inference-time control framework termed Semantically Decoupled Latent Steering (SDLS). Unlike generic activation steering, which often suffers from semantic entanglement, our approach constructs a semantic-free intervention vector via large language model (LLM)-driven semantic decomposition followed by $QR$-based orthogonalization. This orthogonalization step is critical. It leverages geometric constraints to filter out the clinical semantics often entangled in standard principal component analysis (PCA) directions, ensuring that the steering vector targets only the ``historical comparison" axis. We validate our method on the BiomedGPT foundation model, demonstrating that it overcomes the trade-off between hallucination suppression and clinical accuracy. Extensive experiments on MIMIC-CXR, and zero-shot transfer evaluation on CheXpert Plus and IU-Xray, demonstrate the robustness of our approach. Quantitative evaluations on MIMIC-CXR show that our approach significantly reduces the probability of historical hallucinations (FilBERT score decreases from 0.2373 to 0.1889) and improves clinical label fidelity (CheXpert macro-F1 increases from 0.2242 to 0.3208). Supplementary evaluations confirm that the structural integrity of the clinical narrative is maintained.

cs.CV↗

arXiv2026

HumanOrbit: 3D Human Reconstruction as 360° Orbit Generation

We present a method for generating a full 360° orbit video around a person from a single input image. Existing methods typically adapt image-based diffusion models for multi-view synthesis, but yield inconsistent results across views and with the original identity. In contrast, recent video diffusion models have demonstrated their ability in generating photorealistic results that align well with the given prompts. Inspired by these results, we propose HumanOrbit, a video diffusion model for multi-view human image generation. Our approach enables the model to synthesize continuous camera rotations around the subject, producing geometrically consistent novel views while preserving the appearance and identity of the person. Using the generated multi-view frames, we further propose a reconstruction pipeline that recovers a textured mesh of the subject. Experimental results validate the effectiveness of HumanOrbit for multi-view image generation and that the reconstructed 3D models exhibit superior completeness and fidelity compared to those from state-of-the-art baselines.

cs.CV↗

arXiv2026

Crosscap states and duality of Ising field theory in two dimensions

We propose two distinct crosscap states for the two-dimensional (2D) Ising field theory. These two crosscap states, identifying Ising spins or dual spins (domain walls) at antipodal points, are shown to be related via the Kramers-Wannier duality transformation. We derive their Majorana free field representations and extend bosonization techniques to calculate correlation functions of the 2D Ising conformal field theory (CFT) with different crosscap boundaries. Away from criticality, we develop a conformal perturbation theory to calculate the Klein bottle entropy (norm-square of the crosscap overlap) as a universal scaling function [Phys. Rev. Lett. 130, 151602 (2023)]. For the Ising field theory, our analytical results support the conjectured monotonicity of the Klein bottle entropy under relevant perturbations. The formalism provides a general framework for studying perturbed 2D CFTs on non-orientable manifolds.

cond-mat.str-el↗

arXiv2026

Interface Spin-orbit Coupling Induced Room-temperature Ferromagnetic Insulator

Fabricating room-temperature ferromagnetic insulators, which are crucial candidates for next-generation dissipation-free quantum and spintronic devices, remains a significant challenge. In this Letter, we report on the epitaxial synthesis of novel room-temperature ferromagnetic insulating thin films created through the precise construction of (111)-oriented 3d/5d interfaces. Our analysis indicates that, unlike conventional doping methods, the (111)-oriented SrIrO3/La2/3Sr1/3MnO3 (SIO/LSMO) interfaces exhibit markedly enhanced spin-orbit coupling. This enhanced interfacial spin-orbit coupling strengthens the electron-phonon coupling in LSMO, thereby shortening the electronic mean free path. As a result, the intrinsic metallicity of LSMO is suppressed, giving rise to a new ferromagnetic insulating phase that emerges between the ferromagnetic metal and paramagnetic insulator regimes of the LSMO phase diagram. Furthermore, the temperature window of the ferromagnetic insulating phase can be tuned by precisely controlling the thickness of the LSMO layers. Our Letter reveals a new strategy for developing ferromagnetic insulators by engineering 3d/5d interfaces and orientations, paving a way for the development of novel dissipation-free quantum and spintronic devices.

cond-mat.str-el↗

arXiv2026

From Perception to Action: An Interactive Benchmark for Vision Reasoning

Understanding the physical structure is essential for real-world applications such as embodied agents, interactive design, and long-horizon manipulation. Yet, prevailing Vision-Language Model (VLM) evaluations still center on structure-agnostic, single-turn setups (e.g., VQA), which fail to assess agents' ability to reason about how geometry, contact, and support relations jointly constrain what actions are possible in a dynamic environment. To address this gap, we introduce the Causal Hierarchy of Actions and Interactions (CHAIN) benchmark, an interactive 3D, physics-driven testbed designed to evaluate whether models can understand, plan, and execute structured action sequences grounded in physical constraints. CHAIN shifts evaluation from passive perception to active problem solving, spanning tasks such as interlocking mechanical puzzles and 3D stacking and packing. We conduct a comprehensive study of state-of-the-art VLMs and diffusion-based models under unified interactive settings. Our results show that top-performing models still struggle to internalize physical structure and causal constraints, often failing to produce reliable long-horizon plans and cannot robustly translate perceived structure into effective actions. The project is available at https://social-ai-studio.github.io/CHAIN/.

cs.CV↗

arXiv2026

TensorCircuit-NG: A Universal, Composable, and Scalable Platform for Quantum Computing and Quantum Simulation

We present TensorCircuit-NG, a next-generation quantum software platform designed to bridge the gap between quantum physics, artificial intelligence, and high-performance computing. Moving beyond the scope of traditional circuit simulators, TensorCircuit-NG establishes a unified, tensor-native programming paradigm where quantum circuits, tensor networks, and neural networks fuse into a single, end-to-end differentiable computational graph. Built upon industry-standard machine learning backends (JAX, TensorFlow, PyTorch), the framework introduces comprehensive capabilities for approximate circuit simulation, analog dynamics, fermion Gaussian states, qudit systems, and scalable noise modeling. To tackle the exponential complexity of deep quantum circuits, TensorCircuit-NG implements advanced distributed computing strategies, including automated data parallelism and model-parallel tensor network slicing. We validate these capabilities on GPU clusters, demonstrating a near-linear speedup in distributed variational quantum algorithms. TensorCircuit-NG enables flagship applications, including end-to-end QML for CIFAR-100 computer vision, efficient pipelines from quantum states to neural networks via classical shadows, and differentiable optimization of tensor network states for many-body physics.

quant-ph↗

arXiv2026

Roadmap for warm dense matter physics

This roadmap presents the state-of-the-art, current challenges and near future developments anticipated in the thriving field of warm dense matter physics. Originating from strongly coupled plasma physics, high pressure physics and high energy density science, the warm dense matter physics community has recently taken a giant leap forward. This is due to spectacular developments in laser technology, diagnostic capabilities, and computer simulation techniques. Only in the last decade has it become possible to perform accurate enough simulations \& experiments to truly verify theoretical results as well as to reliably design experiments based on predictions. Consequently, this roadmap discusses recent developments and contemporary challenges that are faced by theoretical methods, and experimental techniques needed to create and diagnose warm dense matter. A large part of this roadmap is dedicated to specific warm dense matter systems and applications in astrophysics, inertial confinement fusion and novel material synthesis.

physics.plasm-ph↗

arXiv2026

Truth in the Few: High-Value Data Selection for Efficient Multi-Modal Reasoning

While multi-modal large language models (MLLMs) have made significant progress in complex reasoning tasks via reinforcement learning, it is commonly believed that extensive training data is necessary for improving multi-modal reasoning ability, inevitably leading to data redundancy and substantial computational costs. However, can smaller high-value datasets match or outperform full corpora for multi-modal reasoning in MLLMs? In this work, we challenge this assumption through a key observation: meaningful multi-modal reasoning is triggered by only a sparse subset of training samples, termed cognitive samples, whereas the majority contribute marginally. Building on this insight, we propose a novel data selection paradigm termed Reasoning Activation Potential (RAP)}, which identifies cognitive samples by estimating each sample's potential to stimulate genuine multi-modal reasoning by two complementary estimators: 1) Causal Discrepancy Estimator (CDE) based on the potential outcome model principle, eliminates samples that overly rely on language priors by comparing outputs between multi-modal and text-only inputs; 2) Attention Confidence Estimator (ACE), which exploits token-level self-attention to discard samples dominated by irrelevant but over-emphasized tokens in intermediate reasoning stages. Moreover, we introduce a Difficulty-aware Replacement Module (DRM) to substitute trivial instances with cognitively challenging ones, thereby ensuring complexity for robust multi-modal reasoning. Experiments on six datasets show that our RAP method consistently achieves superior performance using only 9.3% of the training data, while reducing computational costs by over 43%.

cs.CV↗

arXiv2026

MedScope: Incentivizing "Think with Videos" for Clinical Reasoning via Coarse-to-Fine Tool Calling

Long-form clinical videos are central to visual evidence-based decision-making, with growing importance for applications such as surgical robotics and related settings. However, current multimodal large language models typically process videos with passive sampling or weakly grounded inspection, which limits their ability to iteratively locate, verify, and justify predictions with temporally targeted evidence. To close this gap, we propose MedScope, a tool-using clinical video reasoning model that performs coarse-to-fine evidence seeking over long-form procedures. By interleaving intermediate reasoning with targeted tool calls and verification on retrieved observations, MedScope produces more accurate and trustworthy predictions that are explicitly grounded in temporally localized visual evidence. To address the lack of high-fidelity supervision, we build ClinVideoSuite, an evidence-centric, fine-grained clinical video suite. We then optimize MedScope with Grounding-Aware Group Relative Policy Optimization (GA-GRPO), which directly reinforces tool use with grounding-aligned rewards and evidence-weighted advantages. On full and fine-grained video understanding benchmarks, MedScope achieves state-of-the-art performance in both in-domain and out-of-domain evaluations. Our approach illuminates a path toward medical AI agents that can genuinely "think with videos" through tool-integrated reasoning. We will release our code, models, and data.

cs.CV↗

arXiv2026

AgentSkiller: Scaling Generalist Agent Intelligence through Semantically Integrated Cross-Domain Data Synthesis

Large Language Model agents demonstrate potential in solving real-world problems via tools, yet generalist intelligence is bottlenecked by scarce high-quality, long-horizon data. Existing methods collect privacy-constrained API logs or generate scripted interactions lacking diversity, which struggle to produce data requisite for scaling capabilities. We propose AgentSkiller, a fully automated framework synthesizing multi-turn interaction data across realistic, semantically linked domains. It employs a DAG-based architecture with explicit state transitions to ensure determinism and recoverability. The pipeline builds a domain ontology and Person-Centric Entity Graph, defines tool interfaces via Service Blueprints for Model Context Protocol servers, and populates environments with consistent databases and strict Domain Policies. A cross-domain fusion mechanism links services to simulate complex tasks. Finally, the pipeline creates user tasks by verifying solution paths, filtering via execution-based validation, and generating queries using a Persona-based Simulator for automated rollout. This produces reliable environments with clear state changes. To demonstrate effectiveness, we synthesized $\approx$ 11K interaction samples; experimental results indicate that models trained on this dataset achieve significant improvements on function calling over baselines, particularly in larger parameter regimes.

cs.CL↗

arXiv2026

SPPAM: Signature Pattern Prediction and Access-Map Prefetcher

The discrepancy between processor speed and memory system performance continues to limit the performance of many workloads. To address the issue, one effective and well studied technique is cache prefetching. Many prefetching designs have been proposed, with varying approaches and effectiveness. For example, SPP is a popular prefetcher that leverages confidence throttled recursion to speculate on the future path of program's references, however it is very susceptible to the reference reordering of higher-level caches and the out-of-order core. Orthogonally, AMPM is another popular approach to prefetching which uses reordering-resistant access maps to identify patterns within a region, but is unable to speculate beyond that region. In this paper, we propose SPPAM, a new approach to prefetching, inspired by prior works such as SPP and AMPM, while addressing their limitations. SPPAM utilizes online-learning to build a set of access-map patterns. These patterns are used in a speculative lookahead which is throttled by a confidence metric. Targeting the second-level cache, SPPAM alongside state-of-the-art prefetchers Berti and Bingo improves system performance by 31.4% over no prefetching and 6.2% over the baseline of Berti and Pythia.

cs.AR↗

arXiv2026

Complexity of Projected Gradient Methods for Strongly Convex Optimization with Hölder Continuous Gradient Terms

This paper studies the complexity of projected gradient descent methods for a class of strongly convex constrained optimization problems where the objective function is expressed as a summation of $m$ component functions, each possessing a gradient that is Hölder continuous with an exponent $α_i \in (0, 1]$. Under this formulation, the gradient of the objective function may fail to be globally Hölder continuous, thereby rendering existing complexity results inapplicable to this class of problems. Our theoretical analysis reveals that, in this setting, the complexity of projected gradient methods is determined by $\hatα = \min_{i \in \{1, \dotsc, m\}} α_i$. We first prove that, with an appropriately fixed stepsize, the complexity bound for finding an approximate minimizer with a distance to the true minimizer less than $\varepsilon$ is $O (\log (\varepsilon^{-1}) \varepsilon^{2 (\hatα - 1) / (1 + \hatα)})$, which extends the well-known complexity result for $\hatα = 1$. Next we show that the complexity bound can be improved to $O (\log (\varepsilon^{-1}) \varepsilon^{2 (\hatα - 1) / (1 + 3 \hatα)})$ if the stepsize is updated by the universal scheme. We illustrate our complexity results by numerical examples arising from elliptic equations with a non-Lipschitz term.

math.OC↗