arXiv subjects

Shiwei Zhang

Publications and source records attributed to Shiwei Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Signature of Short-Range Order in Static Response of the Three-Dimensional Electron Gas

The three-dimensional electron gas is a fundamental model in condensed matter physics and quantum chemistry, and the exchange-correlation energy derived from it serves as the starting point of \textit{ab initio} computations of materials. However, the wave-vector-dependent response, and hence the static local field factor $G(q)$, has remained without accurate ground-state benchmark in the strongly coupled regime for three decades. Using diffusion Monte Carlo, we calculate the static density-density response function and $G(q)$ across the liquid phase and find a pronounced structure in $G(q)$ at intermediate wave vectors, already visible at metallic densities and growing with increasing interaction strength. We identify it as a fingerprint of short-range order by showing that it is required to reproduce the static structure factor. Our parametrization, valid in the entire liquid phase, predicts a low-energy resonant mode inside the particle-hole continuum.

physics.comp-ph↗

arXiv2026

Static Response of the Strongly Coupled Electron Gas from Diffusion Monte Carlo

The local field factor $G(q)$ of the HEG describes how exchange and correlation modify the response to a static perturbation, and enters dielectric screening, effective electron--electron interactions, and the xc kernel of TDDFT. Accurate zero-temperature data for $G(q)$, however, remain limited to metallic densities, $r_s\leq10$. In this work, we present diffusion Monte Carlo calculations of the static density response function $χ(q)$ and of $G(q)$ for $2\leq r_s\leq100$, and describe and discuss the methodological steps that enabled these calculations. We show how optimization of the perturbing potential in the trial wave function has an important effect, as a poor choice can reverse the sign of the response. A closed-form expression is provided for its optimal value, together with the density dependence of the linear-response window. We further show that backflow correlations leave $χ(q)$ unchanged and that $G(q)$ depends only weakly on system size even at very low density, so that Slater--Jastrow wave functions and modest system sizes suffice. Together, these results provide a practical guide for future QMC calculations of response properties of the electron gas and real materials. Our results agree with previous DMC data at metallic densities for $q>k_F$, and extend them below $k_F$. They also agree with diagrammatic QMC at $r_s=2$. At lower densities, our results show that $G(q)$ develops a nonmonotonic structure at intermediate-$q$, where the response approaches that of the noninteracting gas. Quantitative agreement is seen with finite-temperature path integral Monte Carlo, suggesting that the previously reported finite-temperature structure is inherited from the ground state. Using a parametrization of $G(q)$ that reproduces our data throughout the liquid phase, we relate this structure to short-range order in the pair correlation function and the static structure factor.

cond-mat.str-el↗

arXiv2026

Removing nodal and support-mismatch pathologies in Variational Monte Carlo via blurred sampling

Variational Monte Carlo (VMC) is a powerful and fast-growing method for optimizing and evolving parameterized many-body wave functions, especially with modern neural-network quantum states. In practice, however, the stochastic estimators that form the backbone of the method can become unstable or biased due to the presence of nodes, a ubiquitous feature of quantum wave functions. In the continuum, this results in heavy-tailed estimators with potentially divergent variances, while in discrete Hilbert spaces the sampling distribution can miss parts of the support needed to form unbiased estimators. These statistical pathologies lead to unreliable optimization trajectories in stochastic reconfiguration or incorrect variational dynamics in time-dependent Variational Monte Carlo (t-VMC), and severely limit the power of the numerical simulations. We introduce blurred sampling to address these difficulties. The method has a number of rigorous properties that make it well-behaved, effective and computationally efficient. Additionally, it is a post-processing approach that can be used without modifying the underlying sampler and incurs only minimal overhead. We demonstrate its effectiveness in a broad range of problems -- including ground state optimization in both continuum and discrete settings, as well as in time-dependent variational Monte Carlo, where it resolves previously identified failure modes -- and further apply it to large-scale problems in spin dynamics. This work establishes a broadly applicable framework for robust VMC and t-VMC calculations.

cond-mat.str-el↗

arXiv2026

DiffusionOPD: A Unified Perspective of On-Policy Distillation in Diffusion Models

Reinforcement learning has emerged as a powerful tool for improving diffusion-based text-to-image models, but existing methods are largely limited to single-task optimization. Extending RL to multiple tasks is challenging: joint optimization suffers from cross-task interference and imbalance, while cascade RL is cumbersome and prone to catastrophic forgetting. We propose DiffusionOPD, a new multi-task training paradigm for diffusion models based on Online Policy Distillation (OPD). DiffusionOPD first trains task-specific teachers independently, then distills their capabilities into a unified student along the student own rollout trajectories. This decouples single-task exploration from multi-task integration and avoids the optimization burden of solving all tasks jointly from scratch. Theoretically, we lift the OPD framework from discrete tokens to continuous-state Markov processes, deriving a closed-form per-step KL objective that unifies both stochastic SDE and deterministic ODE refinement via mean-matching. We formally and empirically demonstrate that this analytic gradient provides lower variance and better generality compared to conventional PPO-style policy gradients. Extensive experiments show that DiffusionOPD consistently surpasses both multi-reward RL and cascade RL baselines in training efficiency and final performance, while achieving state-of-the-art results on all evaluated benchmarks.

cs.LG↗

arXiv2026

OmniEdit-Bench: A Comprehensive Benchmark for Instruction-based Video Editing

Instruction-based video editing (IVE) is an emerging field with broad applications, yet evaluating editing models remains challenging. Existing benchmarks suffer from two major limitations: limited task coverage inherited from image editing, which overlooks video-specific dimensions, and inadequate metrics that fail to measure instruction fidelity, allowing incorrect edits to receive high scores due to strong visual priors from the original video. To address these issues, we introduce a comprehensive and structured benchmark for IVE. Our benchmark decomposes editing tasks into multiple video-specific dimensions, including spatial, temporal, audio, and reference-based editing, extending beyond conventional frame-level evaluation. It also distinguishes explicit and implicit instructions and incorporates reasoning-based scenarios to better reflect real-world requirements. Furthermore, we propose an evaluation framework that assesses editing quality from four complementary dimensions: accuracy, preservation, realism, and consistency, using both human judgments and state-of-the-art vision-language models. To emphasize instruction fidelity, we introduce an accuracy-aware penalty mechanism that conditions other scores on accuracy, preventing visually plausible but incorrect edits from receiving inflated evaluations. Extensive experiments on representative open-source and commercial models show that current IVE models remain far from satisfactory. OmniEdit-Bench provides a comprehensive and reliable testbed for evaluating instruction-based video editing and offers insights into future research directions. The project page is https://omniedit-bench.github.io/.

cs.CV↗

arXiv2026

Visualizing the Impact of Quenched Disorder on 2D Electron Wigner Solids

Electron Wigner solids (WSs)1-12 provide an ideal system for understanding the competing effects of electron-electron and electron-disorder interactions, a central unsolved problem in condensed matter physics. Progress in this topic has been limited by a lack of single-defect-resolved experimental measurements as well as accurate theoretical tools to enable realistic experiment/theory comparison. Here we overcome these limitations by combining atomically-resolved scanning tunneling microscopy (STM) with neural-quantum-state quantum Monte Carlo (NQS-QMC) simulation of disordered 2D electron WSs to discover new disorder-induced physical regimes of correlated electron behavior. STM was used to image the electron density ($n_e$) dependent evolution of electron WSs in gate-tunable bilayer MoSe2 devices with varying long-range ($n_\mathrm{LR}$) and short-range ($n_\mathrm{SR}$) disorder densities. These images were compared to NQS-QMC simulations using realistic disorder maps extracted from experiment, thus allowing the roles of different disorder types to be disentangled. We identify two distinct physical regimes for disordered electron WSs that depend on $n_\mathrm{SR}$. For $n_\mathrm{SR} \lesssim n_e$ the WS behavior is dominated by long-range disorder and features extensive mixed solid-liquid phases, a new type of local re-entrant melting/crystallization, and prominent Friedel oscillations. In contrast, when $n_\mathrm{SR} \gg n_e$ these features are suppressed and a more robust amorphous WS phase emerges that persists to higher ne, highlighting the importance of short-range disorder in this regime. Our work establishes a powerful framework for studying disordered quantum solids via a combined experimental-theoretical approach.

cond-mat.str-el↗

arXiv2026

AutoHF: a general Hartree-Fock solver utilizing direct energy minimization with automatic differentiation

We present autohf, a general, easy-to-use mean-field solver for quantum many-fermion Hamiltonians. It allows the user to bypass the process of deciphering the mean-field form for each many-body Hamiltonian $H$ and thus avoid setting up a tailored program for each $H$. Rather, autohf finds the optimal Slater determinant $|Ψ\rangle$, written in terms of orbital coefficients and subject to symmetry constraints, by directly minimizing the variational energy $\langle H \rangle$. By embracing this variational approach, autohf makes use of the growing power of automatic differentiation and optimization tools developed by the machine learning community.

cond-mat.str-el↗

arXiv2026

FilmWorld: Agentic Novel-to-Film Generation through Dynamic Cinematic World Modeling

Translating novels into films poses a grand challenge for generative artificial intelligence, requiring conversion of abstract literary prose into long-form, multi-scene visual narratives. While current video generation models excel at short, single-scene clips within narrow temporal and spatial contexts, novel-to-film generation operates in a more complex regime, demanding long-duration content across diverse scenes with dynamically evolving entity states. To address this, we formalize novel-to-film generation as dynamic cinematic world modeling, decomposed into two phases: construction, which grounds abstract, underspecified literary narratives into concrete, stateful, and persistent world entities; and evolution, which governs how these entities dynamically update under plot progression to maintain causal consistency across scenes. We propose FilmWorld, an end-to-end agentic system where two groups of specialized agents collaborate to instantiate these phases. Construction-side agents perform narrative structured translation, world entity state modeling with visual anchoring, and state-driven shot planning, progressively projecting literary language into a cinematic blueprint. Evolution-side agents perform state-anchored visual generation, cross-shot dynamic state propagation, and closed-loop state verification to maintain causal consistency and visual coherence. To address the evaluation gap in long-form generation, we introduce FilmEval, a systematic evaluation framework that couples a difficulty-graded benchmark of 15 representative novels with an automated protocol of nine objective metrics spanning three dimensions: cinematic presentation, film consistency, and novel fidelity. Experiments demonstrate that FilmWorld consistently outperforms state-of-the-art video generation agent systems, with particularly pronounced improvements in narrative fidelity and cross-scene consistency.

cs.CV↗

arXiv2026

FinanceComplexQA: Benchmarking Agentic Reasoning on Industrial-grade Financial Documents

Agentic Reasoning has become a transformative force in financial analysis due to its ability to integrate large-scale information and generate reliable and accurate content. However, when handling complex real-world problems, different agents still show significant performance variation. In this work, we design Finance-LaTeX SKILL, a skill for synthesizing financial documents with complex layouts based on expert knowledge. Using an agent workflow built on this skill, we generate 2,000 professional financial documents along with 6,000 high-quality question-answer pairs. To evaluate the overall capability of agents, we introduce FinanceComplexQA, a comprehensive open-ended generation benchmark for financial documents that closely resembles real-world scenarios. It contains 2,026 deep research tasks targeting 1009 financial documents. FinanceComplexQA has 8 key features: bilingual support; coverage of six mainstream scenarios and seven tasks; expert-level document reasoning questions; deep research of complex layouts; relatively stable and permanent reference answers; and precise evaluation through an Agent-as-a-Judge with multiple evaluation metrics. Using FinanceComplexQA, we conduct a comprehensive evaluation of leading RAG systems and agentic reasoning tools for financial document QA. Through identifying and analyzing failure cases, we provide an in-depth study of their capabilities in numerical computation, multi-hop reasoning, content summarization, and industry analysis.

cs.CE↗

arXiv2026

Emergent polaronic correlations in doped spin liquids

The interplay between spin and charge degrees of freedom arising from doping a Mott insulating quantum spin liquid (QSL) has been a topic of research for several decades. Calculating properties of these fractionalized metallic states in single-band models are generally restricted to mean-field patron descriptions and small fluctuations around these states, which are insufficient for quantitative comparison of observables to measurements performed in strongly-correlated systems. In this work, we numerically study a class of correlated electronic wavefunctions which support fractionalized spin and charge excitations and which fully take into account gauge fluctuations through the enforcement of local Hilbert space constraints. By optimizing the energy of these wavefunctions against the hole-doped Fermi Hubbard Hamiltonian, we obtain a variational ansatz for describing the low-energy physics of this model. We compare measurements of hole-induced spin-spin correlation functions to measurements taken in low temperature cold-atom simulations of the Hubbard model and find quantitative agreement between the two. In particular, we demonstrate the emergence of magnetic polaron correlations in these metallic states.

cond-mat.str-el↗

arXiv2026

MSAVBench: Towards Comprehensive and Reliable Evaluation of Multi-Shot Audio-Video Generation

Video generation is rapidly evolving from single-shot synthesis to complex multi-shot audio-video (MSAV) narratives to meet real-world demands. However, evaluating such frontier models remains a fundamental challenge. Existing benchmarks are limited in scope and data diversity, and rely on rigid evaluation pipelines, preventing systematic and reliable assessment of modern MSAV models. To bridge these gaps, we introduce MSAVBench, the first comprehensive benchmark and adaptive hybrid evaluation framework for multi-shot audio-video generation. Our benchmark spans four key dimensions, video, audio, shot, and reference, covering diverse task settings, varying shot counts of up to 15, and challenging non-realistic scenarios. Our evaluation framework improves robustness through an adaptive self-correction mechanism for shot segmentation, instance-wise rubrics for subjective metrics, and tool-grounded evidence extraction for complex judgments. Furthermore, MSAVBench achieves high alignment with human judgments, reaching a Spearman rank correlation of 91.5%. Our systematic evaluation of 19 state-of-the-art closed- and open-source models shows that current systems still struggle with director-level control and fine-grained audio-visual synchronization, while modular or agentic generation pipelines offer a promising path toward narrowing the gap between open- and closed-source models. The benchmark data and evaluation code are publicly available at https://github.com/ali-vilab/MSAVBench.

cs.CV↗

arXiv2026

Transformer-based operator learning framework for self-energy in strongly correlated systems

We introduce $Σ$-Attention, a Transformer-based operator-learning framework to address a key computational challenge in correlated materials. Our approach utilizes an Encoder-Only Transformer as an ansatz to approximate the self-energy operator of strongly correlated electronic systems. By creating a batched dataset that combines results from three complementary approaches: many-body perturbation theory, strong-coupling expansion, and exact diagonalization, each effective in specific parameter regimes, $Σ$-Attention is applied to learn a universal approximation for the self-energy operator that is valid across all regimes. This hybrid strategy leverages the strengths of existing methods while relying on the transformer's ability to generalize beyond individual limitations. More importantly, the scalability of the Transformer architecture allows the learned self-energy to be extended to systems with larger sizes, leading to much improved computational scaling. Using the 1D Hubbard model, we demonstrate that $Σ$-Attention can accurately predict the Matsubara Green's function and hence effectively captures the Mott transition at finite temperature. Our framework offers a promising and scalable pathway for studying strongly correlated systems with many possible generalizations.

cond-mat.str-el↗

arXiv2026

M$^\star$: Every Task Deserves Its Own Memory Harness

Large language model agents rely on specialized memory systems to accumulate and reuse knowledge during extended interactions. Recent architectures typically adopt a fixed memory design tailored to specific domains, such as semantic retrieval for conversations or skills reused for coding. However, a memory system optimized for one purpose frequently fails to transfer to others. To address this limitation, we introduce M$^\star$, a method that automatically discovers task-optimized memory harnesses through executable program evolution. Specifically, M$^\star$ models an agent memory system as a memory program written in Python. This program encapsulates the data Schema, the storage Logic, and the agent workflow Instructions. We optimize these components jointly using a reflective code evolution method; this approach employs a population-based search strategy and analyzes evaluation failures to iteratively refine the candidate programs. We evaluate M$^\star$ on four distinct benchmarks spanning conversation, embodied planning, and expert reasoning. Our results demonstrate that M$^\star$ improves performance over existing fixed-memory baselines robustly across all evaluated tasks. Furthermore, the evolved memory programs exhibit structurally distinct processing mechanisms for each domain. This finding indicates that specializing the memory mechanism for a given task explores a broad design space and provides a superior solution compared to general-purpose memory paradigms.

cs.PL↗

arXiv2026

Bridging Brain and Semantics: A Hierarchical Framework for Semantically Enhanced fMRI-to-Video Reconstruction

Reconstructing dynamic visual experiences as videos from functional magnetic resonance imaging (fMRI) is pivotal for advancing the understanding of neural processes. However, current fMRI-to-video reconstruction methods are hindered by a semantic gap between noisy fMRI signals and the rich content of videos, stemming from a reliance on incomplete semantic embeddings that neither capture video-specific cues (e.g., actions) nor integrate prior knowledge. To this end, we draw inspiration from the dual-pathway processing mechanism in human brain and introduce CineNeuron, a novel hierarchical framework for semantically enhanced video reconstruction from fMRI signals with two synergistic stages. First, a bottom-up semantic enrichment stage maps fMRI signals to a rich embedding space that comprehensively captures textual semantics, image contents, action concepts, and object categories. Second, a top-down memory integration stage utilizes the proposed Mixture-of-Memories method to dynamically select relevant "memories" from previously seen data and fuse them with the fMRI embedding to refine the video reconstruction. Extensive experimental results on two fMRI-to-video benchmarks demonstrate that CineNeuron surpasses state-of-the-art methods across various metrics.

cs.CV↗

arXiv2026

AesRM: Improving Video Aesthetics with Expert-Level Feedback

Despite rapid advances in photorealistic video generation, real-world applications such as filmmaking require video aesthetics, e.g., harmonious colors and cinematic lighting, beyond visual fidelity. Prior work on visual aesthetics largely focuses on images, often reducing aesthetics to coarse definitions, e.g., visual pleasure, without a rigorous and systematic evaluation. To improve video aesthetics, we propose a hierarchical rubric that decomposes video aesthetics into three core dimensions, Visual Aesthetics (VA), Visual Fidelity (VF), and Visual Plausibility (VP), with 15 fine-grained criteria, e.g., shot composition. This framework enables a large-scale expert-annotated preference dataset and an evaluation benchmark, AesVideo-Bench, containing about 2500 video pairs with expert annotations on VA, VF, and VP. We then build a family of Video Aesthetic Reward Models (AesRM): AesRM-Base, which directly predicts pairwise preferences on these dimensions to provide efficient post-training rewards, and AesRM-CoT, which additionally generates CoT aligned with all 15 criteria to improve assessment interpretability. Specifically, we train AesRM with a three-stage progressive scheme: (1) Atomic Aesthetic Capability Learning, which strengthens AesRM's recognition of fundamental aesthetic concepts, e.g., accurately identifying centered composition; (2) Cold-Start, aligning the model with structured reasoning protocols; and (3) GRPO, further improving evaluation accuracy. To enhance AesRM-CoT, we additionally propose self-consistency-based CoT synthesis to improve CoT quality and design CoT-based process rewards during GRPO. Extensive experiments show AesRM outperforms baselines on multiple aesthetics benchmarks and is more robust, with lower position bias. Finally, we align Wan2.2 with AesRM and observe clear aesthetic gains over existing aesthetic reward models.

cs.CV↗

arXiv2026

SynMotion: Semantic-Visual Adaptation for Motion Customized Video Generation

Diffusion-based video motion customization facilitates the acquisition of human motion representations from a few video samples, while achieving arbitrary subjects transfer through precise textual conditioning. Existing approaches often rely on semantic-level alignment, expecting the model to learn new motion concepts and combine them with other entities (e.g., ''cats'' or ''dogs'') to produce visually appealing results. However, video data involve complex spatio-temporal patterns, and focusing solely on semantics cause the model to overlook the visual complexity of motion. Conversely, tuning only the visual representation leads to semantic confusion in representing the intended action. To address these limitations, we propose SynMotion, a new motion-customized video generation model that jointly leverages semantic guidance and visual adaptation. At the semantic level, we introduce the dual-embedding semantic comprehension mechanism which disentangles subject and motion representations, allowing the model to learn customized motion features while preserving its generative capabilities for diverse subjects. At the visual level, we integrate parameter-efficient motion adapters into a pre-trained video generation model to enhance motion fidelity and temporal coherence. Furthermore, we introduce a new embedding-specific training strategy which \textbf{alternately optimizes} subject and motion embeddings, supported by the manually constructed Subject Prior Video (SPV) training dataset. This strategy promotes motion specificity while preserving generalization across diverse subjects. Lastly, we introduce MotionBench, a newly curated benchmark with diverse motion patterns. Experimental results across both T2V and I2V settings demonstrate that \method outperforms existing baselines. Project page: https://lucaria-academy.github.io/SynMotion/

cs.CV↗

arXiv2026

Engineering molecular potential energy surfaces using magnetic cavity quantum electrodynamics

We investigate the effects of coupling a quantum-magnetic cavity field to molecules. Our high-precision auxiliary-field quantum Monte Carlo calculations capture the effect of the cavity field in the presence of electron correlations, and their interplay and competition. In H$_2$, we find that a strong enough cavity coupling makes the original bound ground state metastable, along with inverting the singlet-triplet gap. In ring molecules (e.g., H$_n$), the magnetic cavity coupling stabilizes symmetric geometries. As a consequence, open-shell rings such as H$_4$, H$_8$, or C$_4$H$_4$, which would undergo Jahn-Teller distortions outside of the cavity, obtain exotic spin or ring-current polarized, antiaromatic ground states. These effects are enhanced by increasing the molecule concentration inside the cavity. Our results suggest cavity quantum electrodynamics beyond the long-wavelength approximation as a promising avenue for cavity-altered chemistry.

physics.chem-ph↗

arXiv2026

Quantum Charge-4e Superconductivity and Deconfined Pseudocriticality in the Attractive SU(4) Hubbard Model

Unlike conventional charge-2e superconductors, a charge-4e superconductor exhibits long-range coherence of electron quartets rather than Cooper pairs. Clear zero-temperature realizations of charge-4e superconductivity remain rare. Here, we investigate the zero-temperature phase diagram of the attractive SU(4) Hubbard model with numerically exact, large-scale quantum Monte Carlo (QMC) simulations overcoming major technical hurdles. We identify both charge-2e and charge-4e superconducting phases. Upon increasing interaction, charge-2e correlations are suppressed and eventually vanish, while the charge-4e correlations remain robust and converge with system size, signaling the onset of a quartet-condensed phase. Interestingly, across the charge-2e--charge-4e transition, single electrons remain gapped, while charge-2e correlations exhibit a scaling behavior inconsistent with a conventional Landau description. These features are naturally captured by a fractionalized framework in which the physical charge-2e order parameter is a composite field coupled to an emergent non-Abelian gauge structure. We formulate an Sp(4) gauge-Higgs theory that realizes deconfined quantum pseudocriticality between the Higgs (charge-2e) phase and the confined (charge-4e) phase. The Sp(4) gauge-Higgs theory yields pseudocriticality through a fixed-point collision, and its one-loop collision-point exponents quantitatively track the QMC results. Our results establish charge-4e superconductivity as a bona fide zero-temperature phase, provide a simple model for future studies in a numerically exact framework, and reveal an unconventional route to superconducting criticality.

cond-mat.str-el↗