Searcharxiv⌕ Search

arXiv subjects

Xiaoguang Li

Publications and source records attributed to Xiaoguang Li.

At least 19 recordsLinked to original sources
arXiv2026

Non-trivial topological spin-polarized surface states induced nonreciprocal transport in the hourglass semimetal

Quantum materials such as topological semimetals have gained consummate interest due to their exotic band crossings and protected surface states. In particular, Ta$_3$GeTe$_6$ has emerged as a prototypical hourglass Dirac semimetal, as confirmed by both experimental and theoretical analysis. The ARPES confirms the topological hourglass-type band dispersions and Rashba-type band, while Hall measurements reveal a high hole carrier concentration and nonreciprocal transport behavior. These findings are further supported by a phenomenological toy model, which predicts a nonreciprocal resistance, consistent with the observed direction-dependent magnetoresistance. Furthermore, the Raman spectroscopy uncovers two-fold and four-fold phonon symmetries drag of highly anisotropic phonons, which establishes symmetry-dependent resistance. DFT calculations reveal Dirac-like and hourglass Dirac band crossings near the Fermi level for the bulk, with Rashba-type spin splitting induced by the surface inversion-symmetry breaking, leading to spin-polarized surface states. The combined experimental and theoretical results demonstrate that Ta$_3$GeTe$_6$ hosts a non-trivial electronic topology, with the inversion-symmetry breaking at the surface as the driver of this nonreciprocal transport behavior.

cond-mat.mtrl-sci↗
arXiv2026

PrismaCoder: Pushing the Boundaries of Open Multimodal Code Intelligence through Decomposed Rubric Synthesis

Multimodal code intelligence connects the symbolic world of programs with the visual world their execution creates. Interest in this capability is burgeoning, as models are increasingly asked to produce charts, webpages, animations, and semi-structured documents. Recently, proprietary models have pushed far ahead, producing interactive artifacts even from a text description or a screenshot alone. Yet their data and training remain closed, and open progress is held back by the quantity and quality of available data and by the pipelines and methods for producing it. Quality is the harder of the two, since a sample must be correct as a program, well built, and above all visually right once rendered. No existing signal covers all three. Execution often returns just a pass or fail, code similarity does not map to the visual output, and task-specific verifiers do not generalize. To push open progress further, we start from the data side with PrismaForge, a hierarchical data pipeline centered on a synthesized rubric that decomposes these factors into checkable criteria. The same rubric serves as both filter and reward: it curates the supervised corpus, then drives reinforcement learning on visual-coding tasks. Leveraging PrismaForge for data synthesis and filtering, we build PrismaCoder-2M, the most diverse open dataset for multimodal code intelligence and the largest to date. On the model side, we train PrismaCoder-9B and PrismaCoder-35B on this data and release both. They lead the open state of the art by a wide margin and rival leading proprietary systems. Our analyses further show that the proposed rubric-based pipeline improves multimodal coding performance across model backbones. We release our code and data at https://qiushisun.github.io/PrismaCoder-Home/.

cs.CL↗
arXiv2026

Global dynamics near solitary waves for the generalized Boussinesq equation under even-odd perturbations

We study the global dynamics of the generalized Boussinesq equation near standing solitary waves. For even-odd perturbations, Maulen [J. Math. Pures Appl. 177 (2023)] constructed an asymptotically stable center-stable manifold M near the solitary wave Q. We complete this local description by proving that any sufficiently small even-odd perturbation of Q lying outside M either scatters in the energy space as t -> +infinity or blows up in finite time. The proof relies on two main ingredients. First, we classify even-odd solutions with energy below the ground-state energy by identifying two invariant regions, associated with scattering and blow-up, respectively. Second, we establish a one-pass theorem in one dimension: a non-scattering solution cannot re-enter a suitably chosen neighborhood of the solitary wave once it has left that neighborhood. Consequently, the solution remains confined to one of the two invariant regions. Together with the classification above, this yields a dichotomy between scattering and blow-up outside M.

math.AP↗
arXiv2026

Composition-Driven Metal-to-Semiconductor Transition and Enhanced Phonon Transport in B-C substituted Clathrate

Establishing chemical design rules that simultaneously control the electronic structure and thermal transport is a long-sought goal for heat-management and energy materials. Here, we demonstrate that a single B-to-C substitution changes the electron count and simultaneously reconstructs the bonding network and crystal structure, driving a metal-to-semiconductor transition while concurrently enhancing the lattice thermal conductivity. Using density-functional theory (DFT) and machine-learned interatomic potentials (MLIPs), we investigate the electronic structure, lattice dynamics, and phonon thermal transport in cubic BaB3C3 and its B-to-C-substituted tetragonal BaB2C4 structure. The substitution donates one electron per formula unit to the B-C framework, thereby triggering the formation of strong C-C bonds and opening a bandgap of 0.33 eV. The Message Passing Atomic Cluster Expansion (MACE) model reproduces DFT energies and forces, yielding phonon dispersion and lattice thermal conductivity in excellent agreement with DFT benchmarks. Thermodynamically stable cubic BaB3C3 possesses an isotropic lattice thermal conductivity of 7.6 W/mK at 300 K. Substituting B with C hardens the phonon dispersion (evidenced by a frequency upshift from ~800 to ~900 cm^-1) and simultaneously boosts both phonon group velocities and lifetimes. Consequently, the lattice thermal conductivity surges to 18.7 W/mK, representing a remarkable ~2.5-fold enhancement of the in-plane component and a transformation from isotropic to anisotropic behavior, characterized by kappa_x = kappa_y > kappa_z. Our results demonstrate that minimal B-to-C substitution provides a viable strategy to modulate the electronic structure, thereby transforming the crystal from a metal to a semiconductor and concomitantly enhancing phonon thermal transport.

cond-mat.mtrl-sci↗
arXiv2026

Experience Funnel: A State-Policy Alternating Loop for Self-Evolving Agents

Autonomous agents powered by large language models (LLMs) continuously accumulate experience through interaction, creating an opportunity to improve future behavior through self-evolution. A fundamental challenge is how to transform abundant, task-specific interaction experience into reusable model competence without sacrificing the ability to adapt rapidly to newly observed evidence. Explicit textual states, such as skills and agent harnesses, provide fast, human-readable and editable adaptation, but incur persistent dependence on external context; parametric policies provide compact and reusable competence, but are substantially slower to update. We present \textit{Experience Funnel}, a self-evolving framework that couples fast state adaptation with slow policy consolidation in an alternating loop. Interaction trajectories are first distilled into an explicit textual state, where newly acquired experience can be rapidly incorporated and validated. The framework then selectively identifies state-enabled behavior that remains useful across state revisions and consolidates it into the policy through transition-aware distillation. The updated state--policy pair subsequently generates new rollouts, providing fresh evidence for the next round of state adaptation and policy consolidation. Experiments across diverse agent benchmarks show that \textit{Experience Funnel} consistently improves agent capability over state-only evolution and policy-internalization approaches, while progressively converting useful explicit experience into autonomous policy competence.

cs.CL↗
arXiv2026

Induced electromotive force of a thin metal rod in the alternating electromagnetic field of Helmholtz coil: experimental results and theoretical analysis

We apply a thin metal rod (copper rod) as a probe in the alternating magnetic field generated by a Helmholtz coil, and measure the variation of the induced electromotive force (EMF) on the metal rod at different radial positions of Helmholtz coil. Experimental results show that the induced EMF is zero when the center of the metal rod passes through the center of the cylindrical magnetic field inside the Helmholtz coil. When the metal rod is displaced from the center of field to different radial positions, the induced EMF gradually increases from zero, reaches a maximum at a certain position, and then decreases monotonically as the radial distance continues to increase. At the position where the induced EMF reaches its maximum, the metal rod intersects the radial cross-section of the internal magnetic field of the Helmholtz coil at two points, with a small central portion of the rod located inside the Helmholtz coil and the two end portions outside the coil. To explain the experimental phenomena, we construct four simplified models of the magnetic field distribution based on the actual field distribution of the Helmholtz coil to quantitatively investigate the radial variation of the induced EMF along the metal rod. Our theoretical results show that the four models yield similar results and can all qualitatively explain the experimental data curves, particularly reproducing well the variation trend of the induced EMF and the position of the extremum point. The result of piecewise function fitting model is quantitatively in good agreement with the experimental data. This work is also very much helpful and instructive for undergraduate-level students.

physics.class-ph↗
arXiv2026

OptiMAS: Automatically Optimize Multi-Agent System

Automated evolution of Multi-Agent Systems (MAS) holds significant potential for reducing the manual effort required to design and optimize LLM-based agent architectures. However, extant search-based paradigms face a fundamental trade-off, where an expanded optimization scope exacerbates evolutionary instability, while discrete branch-and-discard search isolates insights across lineages. To address these limitations, we propose a continuous, data-driven optimization paradigm built upon a unified ReAct-based infrastructure that reconciles a broad optimization scope with operational stability. Under this paradigm, we present OptiMAS, a task-agnostic agentic optimizer that leverages textual interaction trajectories and task feedback as loss signals for end-to-end MAS evolution. Equipped with a novel dual-track memory mechanism, OptiMAS sustains performance improvement over extended optimization horizons. Evaluation on four heterogeneous agentic benchmarks with three varying scale and accessibility LLM backbones, demonstrates that OptiMAS consistently achieves competitive or superior accuracy relative to both domain-specialized hand-crafted systems and existing evolutionary methods. Our work establishes a practical milestone toward robust, automated MAS evolution.

cs.MA↗
arXiv2026

Beyond Outcome Rewards: Step-Level Self-Distilled Policy Optimization for Deep Search Agents

Deep search agents operate over trajectories spanning dozens of steps, yet standard reinforcement learning provides only a single outcome reward per trajectory, which is far too sparse for effective credit assignment. On-policy self-distillation (OPSD) addresses this by using the model's own logits as dense token-level teachers, but extending it to search agents introduces a fundamental tension: the teacher, having access to privileged information such as the correct answer, produces a distribution that differs systematically from the student's exploration-based reasoning, and naive distillation causes the student to inherit this information asymmetry rather than learn better search strategies. We resolve this tension through two contributions. First, we construct Evidence Anchors, which are concise, step-level evidence snippets extracted from the web, as privileged information that captures key reasoning steps without revealing the entire answer path. Second, we propose Step-Level Self-Distilled Policy Optimization (SSPO), which converts teacher-student disagreement into step-level advantage weights within GRPO, applied exclusively to incorrect trajectories. This design decouples what to update from how much to update: the outcome reward determines the direction of policy change, while the teacher modulates its magnitude at each step. Correct trajectories are left untouched, preserving their diversity. On Qwen3-8B, SSPO consistently outperforms GRPO across BrowseComp, GAIA, and FRAMES, surpassing or matching GRPO trained with twice as many gradient steps while adding only about 5 percent overhead per step from a single additional forward pass.

cs.LG↗
arXiv2026

iSMART: An Iterative Sampling-and-Regression Technique for Solving Martingale-Based PDEs

We propose the {\bf i}terative {\bf S}a{\bf M}pling-{\bf A}nd-{\bf R}egression {\bf T}echnique (iSMART) for high-dimensional martingale-based partial differential equations (PDEs) in this paper. By leveraging the $L^2$-projection property of conditional expectation and adopting the stop-gradient technique, iSMART reformulates the continuous martingale condition derived from PDEs into a sequence of tractable sampling-regression problems within an iterative framework. This approach relies solely on standard SDE path simulation and plain squared-error loss minimization, completely bypassing the need for adversarial optimization or nested expectation estimation in previous methods. iSMART accommodates linear, semi-linear, and fully nonlinear martingale-based PDEs within a unified iterative procedure. In particular, for fully nonlinear Hamilton-Jacobi-Bellman (HJB) equations, a freezing-and-compensating technique is introduced to strategically shift a portion of the nonlinearity into the SDE drift, thereby improving the convergence behavior of the iterations. Numerous numerical experiments on linear reaction-diffusion equations with sharp gradients, semilinear Burgers-type equations, and fully nonlinear HJB equations demonstrate the accuracy, efficiency, and robustness of the proposed approach in various high dimensions.

math.NA↗
arXiv2026

TXL Fusion: A Hybrid Machine Learning Framework Integrating Chemical Heuristics and Large Language Models for Topological Materials Discovery

Topological materials, including topological insulators (TIs) and topological semimetals (TSMs), offer promising platforms for quantum, spintronic, and low-dissipation electronic technologies. Their discovery, however, remains constrained by the high cost of first-principles calculations and the slow, resource-intensive nature of experimental validation. Here, we introduce TXL Fusion, a hybrid machine-learning framework that integrates chemically inspired heuristics, physically interpretable numerical descriptors, and large language model (LLM)-derived semantic embeddings for topological-materials classification and discovery. By combining space-group symmetry, electron-count and orbital descriptors, composition-derived topological heuristics, and physics-aware semantic representations, TXL Fusion classifies materials into trivial, TSM, and TI categories with improved overall performance and enhanced minority-class TI recognition relative to conventional descriptor-based baselines. The model further serves as a high-throughput pre-screening tool for external discovery spaces, rapidly prioritizing candidate TSMs before expensive first-principles or experimental validation. Representative TXL-prioritized candidates were subsequently supported by density functional theory (DFT) calculations, demonstrating the practical value of the framework for reducing discovery cost. By uniting symbolic chemical rules, statistical learning, and language-based representations, TXL Fusion provides a scalable and interpretable strategy for accelerating the discovery of next-generation topological and quantum materials.

cond-mat.mtrl-sci↗
arXiv2026

Quantum-inspired Chemical Rule for Discovering Topological Materials

Topological materials exhibit unique electronic structures that underpin both fundamental quantum phenomena and next-generation technologies, yet their discovery remains constrained by the high computational cost of first-principles calculations and the slow, resource-intensive nature of experimental synthesis. Recent machine-learning approaches, such as the heuristic topogivity rule, offer a data-driven pre-screening tool by quantifying each element's intrinsic tendency toward topological behavior. Here, we develop a hybrid quantum-classical neural network (HQCNN) that extends this rule into a quantum-inspired formulation. Within this framework, the HQCNN maps compositional descriptors to quantum probability amplitudes, naturally introducing pairwise inter-element correlations inaccessible to classical heuristics. The physical validity of these correlations is substantiated by constructing an equivalent complex-valued neural network (CVNN), confirming both the consistency and interpretability of the formulation. Retaining the simplicity of chemical reasoning while embedding quantum-native features, our quantum-inspired rule enables efficient and generalizable topological classification. High-throughput screening combined with first-principles (DFT) validation reveals five previously unreported topological compounds, demonstrating the enhanced predictive power and physical insight afforded by quantum-inspired heuristics.

cond-mat.mtrl-sci↗
arXiv2026

KG2Code: Bridging Knowledge Graphs and Large Language Models via Executable Code for Question Answering

Recent research has explored the integration of knowledge graphs (KGs) with large language models (LLMs) to enhance their performance on downstream knowledge-intensive tasks, particularly knowledge graph question answering (KGQA). Existing approaches primarily combine LLMs with KGs through retrieval-augmented generation (RAG)-based, agent-based, and SPARQL-based methods. Although these methods have achieved notable success, they still suffer from several limitations, including structural information loss, unfaithful reasoning, and limited flexibility and generalization. To address these challenges, this paper proposes KG2Code, a novel approach that transforms knowledge graphs into a code-based representation, preserving structural semantics while naturally aligning with the code-aware pretraining of modern LLMs. Based on KG2Code, KG2Code-QA is further introduced as a KGQA framework that formulates KGQA as a code generation task. This formulation enables the generation of verifiable reasoning traces and executable code, thereby substantially mitigating the impact of hallucinations. In addition, an automated pipeline is developed to construct a large-scale, high-quality code corpus for effectively training open-source LLMs on KG2Code-QA. After training, LLMs are able to perform KGQA in zero-shot scenarios. Extensive experiments demonstrate that the proposed approach significantly outperforms existing KG-enhanced LLM methods for KGQA, while exhibiting strong generalization to unseen KGs. The code and data are available at Github.

cs.AI↗
arXiv2026

MTGenRec: An Efficient Distributed Training System for Generative Recommendation Models in Meituan

Recommendation is crucial for both user experience and company revenue in Meituan as a leading lifestyle company, and generative recommendation models (GRMs) are shown to produce quality recommendations recently. However, existing systems are limited by insufficient functionality support and inefficient implementations for training GRMs in industrial scenarios. As such, we introduce MTGenRec as an efficient and scalable system for GRM training. Specifically, to handle real-time insertions/deletions of sparse embeddings, MTGenRec employs dynamic hash tables to replace static ones. To improve training efficiency, MTGenRec conducts dynamic sequence balancing to address the computation load imbalances among GPUs and adopts feature ID deduplication alongside automatic table merging to accelerate embedding lookup. Extensive experiments show that MTGenRec improves training throughput by $1.6\times -- 2.4\times$ while achieving good scalability when running over 100 GPUs. MTGenRec has been deployed for many applications in Meituan and is now handling hundreds of millions of requests on a daily basis. On the delivery platform, we observe a 1.22% growth in user order volume and a 1.31% enhancement in online PV_CTR.

cs.DC↗
arXiv2026

Sharp Exponent of Stable Standing Waves for the Perturbated Hartree Equation

This paper is concerned with the stability of standing waves for the mass-critical Hartree equation with a focusing perturbation by the variational method. The profile decomposition theory is employed to prove the attainability of the cross constrained variational problem, and then the comparison of two cross constrained variational problems is derived. The sharp criteria of blowup, the orbital stability, and strong instability of standing waves without any frequency constraint are obtained. This improves the cross constrained variational argument proposed by Zhang (2005).

math.AP↗
arXiv2026

DLLM Agent: See Farther, Run Faster

Diffusion large language models (DLLMs) have emerged as an alternative to autoregressive (AR) decoding with appealing efficiency and modeling properties, yet their implications for agentic multi-step decision making remain underexplored. We ask a concrete question: when the generation paradigm is changed but the agent framework and supervision are held fixed, do diffusion backbones induce systematically different planning and tool-use behaviors, and do these differences translate into end-to-end efficiency gains? We study this in a controlled setting by instantiating DLLM and AR backbones within the same agent workflow (DeepDiver) and performing matched agent-oriented fine-tuning on the same trajectory data, yielding diffusion-backed DLLM Agents and directly comparable AR agents. Across benchmarks and case studies, we find that, at comparable accuracy, DLLM Agents are on average over 30% faster end to end than AR agents, with some cases exceeding 8x speedup. Conditioned on correct task completion, DLLM Agents also require fewer interaction rounds and tool invocations, consistent with higher planner hit rates that converge earlier to a correct action path with less backtracking. We further identify two practical considerations for deploying diffusion backbones in tool-using agents. First, naive DLLM policies are more prone to structured tool-call failures, necessitating stronger tool-call-specific training to emit valid schemas and arguments. Second, for multi-turn inputs interleaving context and action spans, diffusion-style span corruption requires aligned attention masking to avoid spurious context-action information flow; without such alignment, performance degrades. Finally, we analyze attention dynamics across workflow stages and observe paradigm-specific coordination patterns, suggesting stronger global planning signals in diffusion-backed agents.

cs.CL↗
arXiv2026

UIS-Digger: Towards Comprehensive Research Agent Systems for Real-world Unindexed Information Seeking

Recent advancements in LLM-based information-seeking agents have achieved record-breaking performance on established benchmarks. However, these agents remain heavily reliant on search-engine-indexed knowledge, leaving a critical blind spot: Unindexed Information Seeking (UIS). This paper identifies and explores the UIS problem, where vital information is not captured by search engine crawlers, such as overlooked content, dynamic webpages, and embedded files. Despite its significance, UIS remains an underexplored challenge. To address this gap, we introduce UIS-QA, the first dedicated UIS benchmark, comprising 110 expert-annotated QA pairs. Notably, even state-of-the-art agents experience a drastic performance drop on UIS-QA (e.g., from 70.90 on GAIA and 46.70 on BrowseComp-zh to 24.55 on UIS-QA), underscoring the severity of the problem. To mitigate this, we propose UIS-Digger, a novel multi-agent framework that incorporates dual-mode browsing and enables simultaneous webpage searching and file parsing. With a relatively small $\sim$30B-parameter backbone LLM optimized using SFT and RFT training strategies, UIS-Digger sets a strong baseline at 27.27\%, outperforming systems integrating sophisticated LLMs such as O3 and GPT-4.1. This demonstrates the importance of proactive interaction with unindexed sources for effective and comprehensive information-seeking. Our work not only uncovers a fundamental limitation in current agent evaluation paradigms but also provides the first toolkit for advancing UIS research, defining a new and promising direction for robust information-seeking systems. The dataset has been released at: https://huggingface.co/datasets/UIS-Digger/UIS-QA.

cs.AI↗
arXiv2026

Alternating Subspace Method for Sparse Recovery of Signals

This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible. Numerous renowned algorithms for tackling the compressed sensing problem employ an alternating strategy, which typically involves data matching in one module and denoising in another. We present a novel approach, the Alternating Subspace Method (ASM), which integrates the principles of the greedy methods (e.g., the orthogonal matching pursuit type methods) and the splitting methods (e.g., the approximate message passing type methods). Crucially, ASM enhances the splitting method by achieving fidelity in a subspace-restricted fashion. \textcolor{black}{We reveal that such a restriction strategy guarantees global convergence via proximal residual control and establish its local geometric convergence on the LASSO problem.} Numerical experiments on the LASSO, channel estimation, and dynamic compressed sensing problems demonstrate its high convergence rate and its capacity to incorporate different prior distributions. Overall, the proposed method is promising in terms of efficiency, accuracy, and flexibility, and has the potential to be competitive in different sparse recovery applications.

cs.IT↗
arXiv2026

Gradually Excavating External Knowledge for Implicit Complex Question Answering

Recently, large language models (LLMs) have gained much attention for the emergence of human-comparable capabilities and huge potential. However, for open-domain implicit question-answering problems, LLMs may not be the ultimate solution due to the reasons of: 1) uncovered or out-of-date domain knowledge, 2) one-shot generation and hence restricted comprehensiveness. To this end, this work proposes a gradual knowledge excavation framework for open-domain complex question answering, where LLMs iteratively and actively acquire external information, and then reason based on acquired historical knowledge. Specifically, during each step of the solving process, the model selects an action to execute, such as querying external knowledge or performing a single logical reasoning step, to gradually progress toward a final answer. Our method can effectively leverage plug-and-play external knowledge and dynamically adjust the strategy for solving complex questions. Evaluated on the StrategyQA dataset, our method achieves 78.17% accuracy with less than 6% parameters of its competitors, setting new SOTA for ~10B-scale LLMs.

cs.CL↗