arXiv subjects

Qiang Xu

Publications and source records attributed to Qiang Xu.

At least 55 records · Page 3Linked to original sources

arXiv2026

Local supersolid in moiré modulated Bose-Hubbard model using density-matrix renormalization group method

The search and characterization of supersolid phases remain a central topic in condensed matter physics. Inspired by the experimental discovery of local superfluid and insulating phases in two-dimensional moiré optical lattices [Meng et al., Nature 615, 231 (2023)], we systematically explore the emergence of a local supersolid ($l$SS) phase in a one-dimensional Bose-Hubbard model subjected to a moiré potential, using the density-matrix renormalization group method. We impose a maximum site occupation $n_{\rm max}=2$ to realize the soft-core boson constraint. In the absence of nearest-neighbor repulsion, we identify the conventional superfluid, local superfluid, Mott insulator, and moiré-induced insulator phases. When the nearest-neighbor repulsion is turned on, the $l$SS phase emerges in the strong-moiré regime. This phase is uniquely characterized by three key signatures: (i) coexisting local staggered density order and local off-diagonal coherence within isolated moiré supercells; (ii) exponentially decaying global off-diagonal correlations; and (iii) a vanishing global structure factor in the thermodynamic limit, while the local structure factor remains finite. These features clearly distinguish the $l$SS from the conventional global supersolid (SS) phase, which exhibits algebraic correlations and a finite global structure factor. Our results provide a complete microscopic picture of local quantum phases in moiré lattices and offer clear experimental observables for detecting $l$SS states with ultracold atoms.

cond-mat.quant-gas↗

arXiv2026

FAME: Forecasting Academic Impact via Continuous-Time Manifold Evolution

Large Language Models (LLMs) are increasingly used to brainstorm and evaluate research ideas, yet assessing such judgments is fundamentally difficult because the true impact of a new idea may take years to emerge. We address this challenge by using the impact forecasting of human-authored manuscripts as a verifiable proxy task. In a prospective forecasting study, we find that frontier LLMs fail to reliably distinguish high-impact papers from ordinary publications, suggesting that static text-based judging is insufficient for scientific evaluation. To address this limitation, we propose $\textbf{FAME}$ ($\underline{\text{F}}$orecasting $\underline{\text{A}}$cademic Impact via Continuous-Time $\underline{\text{M}}$anifold $\underline{\text{E}}$volution), a spatiotemporal framework for modeling the dynamic trajectories of scientific topics. FAME projects papers into a dynamic latent space informed by textual features and a verified knowledge-flow graph, learning geometric constraints that align impactful manuscripts with the forward momentum of their fields. Experiments on 3,200 arXiv papers across three fast-evolving subfields show that FAME consistently and substantially outperforms state-of-the-art LLM evaluators in prospective multidimensional impact forecasting. Furthermore, integrating FAME's dynamic geometric signals into LLMs significantly improves their forecasting performance. These results support manuscript impact forecasting as a useful, measurable proxy benchmark and position FAME as a strong, trajectory-aware foundation for automated scientific evaluation.

cs.LG↗

arXiv2026

MappingEvolve: LLM-Driven Code Evolution for Technology Mapping

Technology mapping is a critical yet challenging stage in logic synthesis. While Large Language Models (LLMs) have been applied to generate optimization scripts, their potential for core algorithm enhancement remains untapped. We introduce MappingEvolve, an open-source framework that pioneers the use of LLMs to directly evolve technology mapping code. Our method abstracts the mapping process into distinct optimization operators and employs a hierarchical agent-based architecture, comprising a Planner, Evolver, and Evaluator, to guide the evolutionary search. This structured approach enables strategic and effective code modifications. Experiments show our method significantly outperforms direct evolution and strong baselines, achieving 10.04\% area reduction versus ABC and 7.93\% versus mockturtle, with 46.6\%--96.0\% $S_{overall}$ improvement on EPFL benchmarks, while explicitly navigating the area--delay trade-off. Our code and data are available at https://github.com/Flians/MappingEvolve.

cs.CE↗

arXiv2026

EditVerse: Unifying Image and Video Editing and Generation with In-Context Learning

Recent advances in foundation models highlight a clear trend toward unification and scaling, showing emergent capabilities across diverse domains. While image generation and editing have rapidly transitioned from task-specific to unified frameworks, video generation and editing remain fragmented due to architectural limitations and data scarcity. In this work, we introduce EditVerse, a unified framework for image and video generation and editing within a single model. By representing all modalities, i.e., text, image, and video, as a unified token sequence, EditVerse leverages self-attention to achieve robust in-context learning, natural cross-modal knowledge transfer, and flexible handling of inputs and outputs with arbitrary resolutions and durations. To address the lack of video editing training data, we design a scalable data pipeline that curates 232K video editing samples and combines them with large-scale image and video datasets for joint training. Furthermore, we present EditVerseBench, the first benchmark for instruction-based video editing covering diverse tasks and resolutions. Extensive experiments and user studies demonstrate that EditVerse achieves state-of-the-art performance, surpassing existing open-source and commercial models, while exhibiting emergent editing and generation abilities across modalities.

cs.CV↗

arXiv2026

EquivFusion: Unifying Hardware Equivalence Checking from Algorithms to Netlists via MLIR

Ensuring functional consistency between high-level algorithmic models and low-level hardware implementations is a critical challenge, particularly as modern design flows increasingly span heterogeneous abstractions--from deep learning frameworks to hardware netlists. In this paper, we present EquivFusion, an end-to-end equivalence checking tool tailored for multi-modal circuit designs. Unlike traditional flows that rely on siloed tools or ad-hoc translation, EquivFusion leverages a verification-oriented MLIR lowering pipeline to unify diverse entry points, including PyTorch, C/C++, Chisel, Verilog, and gate-level netlists, into a common intermediate representation. This architecture enables automated, pairwise equivalence checking across diverse abstraction levels by rigorously translating designs into standard formal verification formats, i.e., SMT-LIB, BTOR2, AIGER. We demonstrate EquivFusion's feasibility to bridge the semantic gap between software specifications and hardware realizations, showcasing its effectiveness in facilitating "shift-left" formal verification for datapath-intensive hardware designs.

cs.AR↗

arXiv2026

Potential-Barrier Affinity Effect in Solid Systems

Electron accumulation in interatomic regions is a fundamental quantum phenomenon dictating chemical bonding and material properties, yet its origin remains elusive across disciplines. Here, we report a quantum accumulation effect -- potential-barrier affinity (PBA) -- revealed by solving the Schrödinger equation for a crystalline potential. PBA effect drives significant interatomic electron accumulation when electron energy exceeds the barrier maximum. This effect essentially enhances interatomic electron density, governing microstructures and properties of condensed matter. Our theory overturns the traditional wisdom that the interstitial electron localization in electride requires potential-well constraints or hybrid orbitals, and it serves as the fundamental mechanism underlying the formation of conventional solid bonding. This work delivers a paradigm shift in understanding electron distribution and establishes a theoretical foundation for the microscopic design of material properties.

cond-mat.mtrl-sci↗

arXiv2026

DiffHLS: Differential Learning for High-Level Synthesis QoR Prediction with GNNs and LLM Code Embeddings

High-Level Synthesis (HLS) compiles C/C++ into RTL, but exploring pragma-driven optimization choices remains expensive because each design point requires time-consuming synthesis. We propose \textbf{\DiffHLS}, a differential learning framework for HLS Quality-of-Result (QoR) prediction that learns from kernel--design pairs: a kernel baseline and a pragma-inserted design variant. \DiffHLS~encodes kernel and design intermediate-representation graphs with dedicated graph neural network (GNN) branches, and augments the delta pathway with code embeddings from a pretrained code large language model (LLM). Instead of regressing absolute targets directly, we jointly predict the kernel baseline and the design-induced delta, and compose them to obtain the design prediction. On PolyBench, \DiffHLS~attains lower average MAPE than GNN baselines under four GNN backbones, and LLM code embeddings consistently improve over a GNN-only ablation. We further validate scalability on the ForgeHLS dataset.

cs.LG↗

arXiv2026

Ming-Flash-Omni: A Sparse, Unified Architecture for Multimodal Perception and Generation

We propose Ming-Flash-Omni, an upgraded version of Ming-Omni, built upon a sparser Mixture-of-Experts (MoE) variant of Ling-Flash-2.0 with 100 billion total parameters, of which only 6.1 billion are active per token. This architecture enables highly efficient scaling (dramatically improving computational efficiency while significantly expanding model capacity) and empowers stronger unified multimodal intelligence across vision, speech, and language, representing a key step toward Artificial General Intelligence (AGI). Compared to its predecessor, the upgraded version exhibits substantial improvements across multimodal understanding and generation. Notably, it achieves strong performance on vision-language understanding benchmarks, with overall scores on par with Gemini 2.5 Pro, and enables seamless switching among multimodal tasks in multi-turn interactions. In speech, it achieves strong performance in contextual and dialect-aware ASR while enabling joint, continuous-generation of speech, sound, and music. In vision, it introduces generative semantic segmentation that achieves competitive standalone performance and enhances spatial control and editing consistency, alongside marked improvements in identity preservation, and high-fidelity in-image text rendering. Together, these capabilities demonstrate that a single unified model can serve as a practical foundation for general-purpose multimodal intelligence.

cs.CV↗

arXiv2026

FBCIR: Balancing Cross-Modal Focuses in Composed Image Retrieval

Composed image retrieval (CIR) requires multi-modal models to jointly reason over visual content and semantic modifications presented in text-image input pairs. While current CIR models achieve strong performance on common benchmark cases, their accuracies often degrades in more challenging scenarios where negative candidates are semantically aligned with the query image or text. In this paper, we attribute this degradation to focus imbalances, where models disproportionately attend to one modality while neglecting the other. To validate this claim, we propose FBCIR, a multi-modal focus interpretation method that identifies the most crucial visual and textual input components to a model's retrieval decisions. Using FBCIR, we report that focus imbalances are prevalent in existing CIR models, especially under hard negative settings. Building on the analyses, we further propose a CIR data augmentation workflow that facilitates existing CIR datasets with curated hard negatives designed to encourage balanced cross-modal reasoning. Extensive experiments across multiple CIR models demonstrate that the proposed augmentation consistently improves performance in challenging cases, while maintaining their capabilities on standard benchmarks. Together, our interpretation method and data augmentation workflow provide a new perspective on CIR model diagnosis and robustness improvements.

cs.CV↗

arXiv2026

Stabilizing Reinforcement Learning for Diffusion Language Models

Group Relative Policy Optimization (GRPO) is highly effective for post-training autoregressive (AR) language models, yet its direct application to diffusion large language models (dLLMs) often triggers reward collapse. We identify two sources of incompatibility. First, GRPO relies on importance ratios defined by sequence probabilities, which are intractable in dLLMs and must be estimated (e.g., via ELBO-based or mean-field likelihood proxies), yielding inherently noisy ratios. Second, standard GRPO's formulation is not designed for estimated ratios: its conditional clipping can be anomalously bypassed by model-agnostic estimation noise, producing gradient spikes, while its fixed group-size normalization amplifies gradient-magnitude fluctuations under high-variance ratio estimates. We show these effects form a self-reinforcing instability loop that drives policy drift and further increases ratio variance. To break this loop, we propose StableDRL, a reformulation of GRPO tailored for dLLMs that uses (i) unconditional clipping to suppress outlier-induced spikes and (ii) self-normalization to constrain updates within the convex hull of per-sample gradients. We further extend StableDRL to block-wise diffusion models via a staircase attention mechanism.

cs.LG↗

arXiv2026

Rethinking Clause Management for CDCL SAT Solvers

Boolean Satisfiability (SAT) solving underpins a wide range of applications in Electronic Design Automation (EDA), particularly formal verification. However, this paper observes that the mainstream clause reduction heuristic in modern SAT solvers becomes ineffective in the critical domain of complex arithmetic circuit verification, such as multipliers. On these instances, the dominant Literal Block Distance (LBD) metric for measuring clause quality degrades into a simple value of clause length, without any perception of dynamic clause usage during solving. To address this issue, a novel clause reduction mechanism is proposed, which is entirely independent of LBD. Its core idea is to decouple and handle separately the two most fundamental characteristics of learnt clauses--inherent lineage and dynamic usage patterns--thereby avoiding the efficiency degradation caused by inappropriately mixing these properties. Experiments show that our method consistently improves mainstream solvers and achieves speedups of up to 5.74x on complex arithmetic circuit problems, while maintaining comparable performance on general-purpose benchmarks. These results challenge the prevailing LBD-centric clause quality metric for clause management.

cs.LO↗

arXiv2026

FormalRTL: Verified RTL Synthesis at Scale

Large language models (LLMs) have demonstrated significant potential in automating hardware synthesis, yet substantial barriers remain for industrial-scale, datapath-centric designs due to ambiguous specifications and a lack of formal correctness guarantees. In this work, we present FormalRTL, a novel end-to-end multi-agent framework that systematically integrates software reference models as formal, executable specifications to guide register-transfer level (RTL) code generation and verification. By tightly coupling planning, synthesis, and formal equivalence checking, FormalRTL achieves scalable and reliable hardware code generation that addresses the critical challenges faced in industrial contexts. The comprehensive evaluation of a new suite of complex industrial-grade benchmarks demonstrates the effectiveness and robustness of our approach. We will open-source the FormalRTL framework and the benchmark suite to facilitate future research in this area.

cs.AR↗

arXiv2026

From Samples to Scenarios: A New Paradigm for Probabilistic Forecasting

Most state-of-the-art probabilistic time series forecasting models rely on sampling to represent future uncertainty. However, this paradigm suffers from inherent limitations, such as lacking explicit probabilities, inadequate coverage, and high computational costs. In this work, we introduce \textbf{Probabilistic Scenarios}, an alternative paradigm designed to address the limitations of sampling. It operates by directly producing a finite set of \{Scenario, Probability\} pairs, thus avoiding Monte Carlo-like approximation. To validate this paradigm, we propose \textbf{TimePrism}, a simple model composed of only three parallel linear layers. Surprisingly, TimePrism achieves 9 out of 10 state-of-the-art results across five benchmark datasets on two metrics. The effectiveness of our paradigm comes from a fundamental reframing of the learning objective. Instead of modeling an entire continuous probability space, the model learns to represent a set of plausible scenarios and corresponding probabilities. Our work demonstrates the potential of the Probabilistic Scenarios paradigm, opening a promising research direction in forecasting beyond sampling.

cs.LG↗

arXiv2026

Making Slow Thinking Faster: Compressing LLM Chain-of-Thought via Step Entropy

Large Language Models (LLMs) using Chain-of-Thought (CoT) prompting excel at complex reasoning but generate verbose thought processes with considerable redundancy, leading to increased inference costs and reduced efficiency. We introduce a novel CoT compression framework based on step entropy, a metric that quantifies \emph{the informational contribution of individual reasoning steps} to identify redundancy. Through theoretical analysis and extensive empirical validation on mathematical reasoning benchmarks, we demonstrate that steps with low entropy are indeed highly redundant. Our experiments reveal that an astonishing 80\% of low-entropy intermediate steps can be pruned with minor degradation in the final answer accuracy across DeepSeek-R1-7B, 14B and Qwen3-8B. This finding sharply contrasts with random or high-entropy pruning, which severely impairs reasoning performance. Building on this, we propose a novel two-stage training strategy combining Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO) reinforcement learning. This approach enables LLMs to autonomously learn to generate compressed COTs during inference by strategically incorporating [SKIP] tokens. Our method significantly improves LLM inference efficiency while preserving accuracy, paving the way for more scalable LLM deployments and a better understanding of their internal reasoning. The code and data are released in https://github.com/staymylove/COT_Compresstion_via_Step_entropy.

cs.AI↗

arXiv2026

Boundary estimates and Green function's expansion for elliptic systems with random coefficients

We investigate boundary estimates for elliptic operators with stationary random coefficients exhibiting integrable correlations, arising from stochastic homogenization theory. As practical applications, we establish decay estimates for Green functions in both quenched and annealed senses. Furthermore, we derive notable annealed estimates for boundary correctors, including central limit theorem (CLT)-scaling type estimates. By extending the lemma of Bella, Giunti, and Otto [10] to accommodate boundary conditions, we ultimately obtain error estimates for the two-scale expansion of Green functions at the level of mixed derivatives, thereby establishing connections to other related fields.

math.AP↗

arXiv2026

CktEvo: Repository-Level RTL Code Benchmark for Design Evolution

Register-Transfer Level (RTL) coding is an iterative, repository-scale process in which Power, Performance, and Area (PPA) emerge from interactions across many files and the downstream toolchain. While large language models (LLMs) have recently been applied to hardware design, most efforts focus on generation or debugging from natural-language prompts, where ambiguity and hallucinations necessitate expert review. A separate line of work begins from formal inputs, yet typically optimizes high-level synthesis or isolated modules and remains decoupled from cross-file dependencies. In this work, we present CktEvo, a benchmark and reference framework for repo-level RTL evolution. Unlike prior benchmarks consisting of isolated snippets, our benchmark targets complete IP cores where PPA emerges from cross-file dependencies. Our benchmark packages several high-quality Verilog repositories from real-world designs. We formalize the task as: given an initial repository, produce edits that preserve functional behavior while improving PPA. We also provide a closed-loop framework that couples LLM-proposed edits with toolchain feedback to enable cross-file modifications and iterative repair at repository scale. Our experiments demonstrate that the reference framework realizes PPA improvements without any human interactions. CktEvo establishes a rigorous and executable foundation for studying LLM-assisted RTL optimization that matters for engineering practice: repository-level, function-preserving, and PPA-driven.

cs.AR↗

arXiv2026

Latent Context Compilation: Distilling Long Context into Compact Portable Memory

Efficient long-context LLM deployment is stalled by a dichotomy between amortized compression, which struggles with out-of-distribution generalization, and Test-Time Training, which incurs prohibitive synthetic data costs and requires modifying model weights, creating stateful parameters that complicate concurrent serving. We propose Latent Context Compilation, a framework that fundamentally shifts context processing from adaptation to compilation. By utilizing a disposable LoRA module as a compiler, we distill long contexts into compact buffer tokens -- stateless, portable memory artifacts that are plug-and-play compatible with frozen base models. Crucially, we introduce a self-aligned optimization strategy that eliminates the need for synthetic context-relevant QA pairs. By regularizing context reconstruction task with context-agnostic random queries, we force compressed tokens to reside within the model's existing instruction-following manifold. Experiments with Llama-3.1-8B demonstrate that Latent Context Compilation preserves fine-grained details and reasoning capabilities where prior methods falter, effectively decoupling memory density from model parameters even at a 16x compression ratio.

cs.LG↗

arXiv2026

Unlocking the Power of Orbital-Free Density Functional Theory to Explore the Electronic Structure Under Extreme Conditions

Recent advances in X-ray free-electron laser diagnostics have enabled direct probing of the electronic structure under extreme pressures and temperatures, such as those encountered in stellar interiors and inertial confinement fusion experiments, challenging theoretical models for interpreting experimental data. Kohn-Sham density functional theory (KSDFT) has been successfully applied to analyze experimental X-ray scattering measurements, but its high computational cost renders routine application impractical. Orbital-free DFT (OFDFT) is a substantially more efficient alternative, with computational cost scaling linearly with system size and a weak temperature dependence, yet it often lacks the accuracy required for electronic structure description. Overcoming this limitation, we present a non-empirical Kohn-Sham-assisted orbital-free density functional framework for calculations at extreme conditions, which enables efficient OFDFT simulations with KSDFT-level accuracy for electron densities, electron-ion structure factors, and equations of state across a broad range of conditions. Benchmark comparisons with quantum Monte Carlo data for dense hydrogen and validation against Rayleigh weight measurements of hot dense beryllium demonstrate the reliability of the framework and speedups of up to several hundred times compared with KSDFT. We further show that even at temperatures on the order of 100 eV, quantum non-locality remains essential for correctly describing the electronic structure of dense hydrogen.

cond-mat.mtrl-sci↗