arXiv subjects

Yi Lu

Publications and source records attributed to Yi Lu.

At least 73 records · Page 4Linked to original sources

arXiv2026

Revisiting the first-order QCD phase transition in dense strong interaction matter

We revisit the phase structure and thermodynamics of QCD in the low temperature and high density region, where a strong, first-order phase transition is expected beyond the critical end point. By solving the quark gap equation in the continuum QCD approach, we reveal the coexistence of the multi-phases both in the microscopic dynamics of chiral symmetry breaking and also in the thermodynamic observables, which manifests the existence of spinodal decomposition during the first-order QCD phase transitions. We also analyse the interface structure of the co-exist Nambu and Wigner phases in the isothermal process during the first-order transition. In particular, the interface tension and interface entropy density are extracted from the isothermal trajectories, which further allows for an analysis on the formation of nuclear bubble, including the bubble radius and its stability at different temperatures. Our predictions may serve as useful inputs for further investigations in heavy-ion physics or astrophysics research.

hep-ph↗

arXiv2026

OPE: Overcoming Information Saturation in Parallel Thinking via Outline-Guided Path Exploration

Parallel thinking has emerged as a new paradigm for large reasoning models (LRMs) in tackling complex problems. Recent methods leverage Reinforcement Learning (RL) to enhance parallel thinking, aiming to address the limitations in computational resources and effectiveness encountered with supervised fine-tuning. However, most existing studies primarily focus on optimizing the aggregation phase, with limited attention to the path exploration stage. In this paper, we theoretically analyze the optimization of parallel thinking under the Reinforcement Learning with Verifiable Rewards (RLVR) setting, and identify that the mutual information bottleneck among exploration paths fundamentally restricts overall performance. To address this, we propose Outline-Guided Path Exploration (OPE), which explicitly partitions the solution space by generating diverse reasoning outlines prior to parallel path reasoning, thereby reducing information redundancy and improving the diversity of information captured across exploration paths. We implement OPE with an iterative RL strategy that optimizes outline planning and outline-guided reasoning independently. Extensive experiments across multiple challenging mathematical benchmarks demonstrate that OPE effectively improves reasoning performance in different aggregation strategies, enabling LRMs to more reliably discover correct solutions.

cs.AI↗

arXiv2026

Dr. Bench: A Multidimensional Evaluation for Deep Research Agents, from Answers to Reports

As an embodiment of intelligence evolution toward interconnected architectures, Deep Research Agents (DRAs) systematically exhibit the capabilities in task decomposition, cross-source retrieval, multi-stage reasoning, information integration, and structured output, which markedly enhance performance on complex and open-ended tasks. However, existing benchmarks remain deficient in evaluation dimensions, response format, and scoring mechanisms, limiting their effectiveness in assessing such agents. This paper introduces Dr. Bench, a multidimensional evaluation framework tailored to DRAs and long-form report-style responses. The benchmark comprises 214 expert-curated challenging tasks across 10 broad domains, each accompanied by manually constructed reference bundles to support composite evaluation. This framework incorporates metrics for semantic quality, topical focus, and retrieval trustworthiness, enabling a comprehensive evaluation of long reports generated by DRAs. Extensive experimentation confirms the superior performance of mainstream DRAs over web-search-tool-augmented reasoning models, yet reveals considerable scope for further improvement. This study provides a robust foundation for capability assessment, architectural refinement, and paradigm advancement of DRAs.

cs.AI↗

arXiv2025

Syzygy of Thoughts: Improving LLM CoT with the Minimal Free Resolution

Chain-of-Thought (CoT) prompting enhances the reasoning of large language models (LLMs) by decomposing problems into sequential steps, mimicking human logic and reducing errors. However, complex tasks with vast solution spaces and vague constraints often exceed the capacity of a single reasoning chain. Inspired by Minimal Free Resolution (MFR) in commutative algebra and algebraic geometry, we propose Syzygy of Thoughts (SoT)-a novel framework that extends CoT by introducing auxiliary, interrelated reasoning paths. SoT captures deeper logical dependencies, enabling more robust and structured problem-solving. MFR decomposes a module into a sequence of free modules with minimal rank, providing a structured analytical approach to complex systems. This method introduces the concepts of "Module", "Betti numbers","Freeness", "Mapping", "Exactness" and "Minimality", enabling the systematic decomposition of the original complex problem into logically complete minimal subproblems while preserving key problem features and reducing reasoning length. We tested SoT across diverse datasets (e.g., GSM8K, MATH) and models (e.g., GPT-4o-mini, Qwen2.5), achieving inference accuracy that matches or surpasses mainstream CoTs standards. Additionally, by aligning the sampling process with algebraic constraints, our approach enhances the scalability of inference time in LLMs, ensuring both transparent reasoning and high performance. Our code will be publicly available at https://github.com/dlMARiA/Syzygy-of-thoughts.

cs.CL↗

arXiv2025

HLTCOE Evaluation Team at TREC 2025: VQA Track

The HLTCOE Evaluation team participated in TREC VQA's Answer Generation (AG) task, for which we developed a listwise learning framework that aims to improve semantic precision and ranking consistency in answer generation. Given a video-question pair, a base multimodal model first generates multiple candidate answers, which are then reranked using a model trained with a novel Masked Pointer Cross-Entropy Loss with Rank Weights. This objective integrates pointer-based candidate selection, rank-dependent weighting, and masked cross-entropy under vocabulary restriction, enabling stable and interpretable listwise optimization. By bridging generative modeling with discriminative ranking, our method produces coherent, fine-grained answer lists. Experiments reveal consistent gains in accuracy and ranking stability, especially for questions requiring temporal reasoning and semantic disambiguation.

cs.CV↗

arXiv2025

Highly Anisotropic Charge Dynamics and Spectral Weight Redistribution in the Trilayer Nickelate La$_{4}$Ni$_{3}$O$_{10}$

We study the $ab$-plane and $c$-axis charge dynamics of La$_{4}$Ni$_{3}$O$_{10}$ using optical spectroscopy. While a pronounced Drude profile, i.e. metallic response, is observed in the $ab$-plane optical conductivity $σ_{1}^{ab}(ω)$, the $c$-axis optical spectra $σ_{1}^{c}(ω)$ exhibit semiconducting behavior. The zero-frequency extrapolation of the optical conductivity $σ_{1}(ω\rightarrow 0) \equiv 1/ρ_{\text{dc}}$ gives a resistivity anisotropy of $ρ_{c}/ρ_{ab} \simeq 366$ at 300~K for La$_{4}$Ni$_{3}$O$_{10}$, which is much larger than the values in iron-based superconductors but comparable to those in high-$T_{c}$ cuprates. The interband response is also highly anisotropic, showing salient orbital selectivity for light polarized in the $ab$ plane and along the $c$ axis. The interband-transition peaks in both $σ_{1}^{ab}(ω)$ and $σ_{1}^{c}(ω)$ are located at lower energies compared to density-functional-theory predictions, signifying considerable electronic correlations. By investigating the spectral weight transfer, we find that in the pristine phase, Coulomb correlations have a marked impact on the charge dynamics of \LNO, whereas in the density-wave state, a gap opens with the Ni-$d_{z^{2}}$ orbital being involved.

cond-mat.supr-con↗

arXiv2025

Dual instability of superconductivity from oxygen defects in La$_3$Ni$_2$O$_{7+δ}$

We uncover a dual mechanism by which oxygen defects suppress superconductivity in the bilayer nickelate La$_3$Ni$_2$O$_{7+δ}$ using density functional theory, dynamical mean-field theory, and functional renormalization group analysis. Apical vacancies and interbilayer interstitials emerge as the dominant low-energy defect species and are further stabilized by orthorhombic domain walls. These two defect classes drive the electronic structure in opposing directions. Vacancy-induced disorder generates local magnetic moments and promotes Anderson localization at moderate concentrations, whereas periodic interstitial ordering yields a coherent but weakly correlated metallic background that fails to support superconductivity. These findings highlight the decisive role of oxygen defects in shaping the superconducting and provide microscopic guidance for improving superconductivity through controlled defect engineering.

cond-mat.supr-con↗

arXiv2025

Breaking Forgetting: Training-Free Few-Shot Class-Incremental Learning via Conditional Diffusion

Efforts to overcome catastrophic forgetting in Few-Shot Class-Incremental Learning (FSCIL) have primarily focused on developing more effective gradient-based optimization strategies. In contrast, little attention has been paid to the training cost explosion that inevitably arises as the number of novel classes increases, a consequence of relying on gradient learning even under extreme data scarcity. More critically, since FSCIL typically provides only a few samples for each new class, gradient-based updates not only induce severe catastrophic forgetting on base classes but also hinder adaptation to novel ones. This paper seeks to break this long-standing limitation by asking: Can we design a training-free FSCIL paradigm that entirely removes gradient optimization? We provide an affirmative answer by uncovering an intriguing connection between gradient-based optimization and the Conditional Diffusion process. Building on this observation, we propose a Conditional Diffusion-driven FSCIL (CD-FSCIL) framework that substitutes the conventional gradient update process with a diffusion-based generative transition, enabling training-free incremental adaptation while effectively mitigating forgetting. Furthermore, to enhance representation under few-shot constraints, we introduce a multimodal learning strategy that integrates visual features with natural language descriptions automatically generated by Large Language Models (LLMs). This synergy substantially alleviates the sample scarcity issue and improves generalization across novel classes. Extensive experiments on mainstream FSCIL benchmarks demonstrate that our method not only achieves state-of-the-art performance but also drastically reduces computational and memory overhead, marking a paradigm shift toward training-free continual adaptation.

cs.CV↗

arXiv2025

Neural network impurity solver for real-frequency dynamical mean-field theory

We introduce a neural network impurity solver for real-frequency DMFT that employs a multihead cross-attention mechanism to map hybridization functions to spectral functions, conditioned on impurity parameters. Trained on high-quality MPS data from complex contour time evolution and incorporating derivative constraints with respect to the complex-time angle, our model achieves smooth generalization to the real-frequency axis. Benchmarking on the single-band Hubbard model for the Bethe lattice demonstrates quantitative accuracy across metallic, strongly correlated, and insulating regimes.

cond-mat.str-el↗

arXiv2025

Preliminary Demonstration of Diamond-GaN pn Diodes via Grafting

Ultrawide bandgap (UWBG) semiconductors exhibit exceptional electrical and thermal properties, offering strong potential for high power and high frequency electronics. However, efficient doping in UWBG materials is typically limited to either n type or p type, constraining their application to unipolar devices. The realization of pn junctions through heterogeneous integration of complementary UWBG or WBG semiconductors is hindered by lattice mismatch and thermal expansion differences. Here, we report the preliminary demonstration of diamond GaN heterojunction pn diodes fabricated via grafting. A single crystalline p plus diamond nanomembrane was integrated onto an epitaxially grown c plane n plus GaN substrate with an ultrathin ALD Al2O3 interlayer. The resulting diodes exhibit an ideality factor of 1.55 and a rectification ratio of over 1e4. Structural and interfacial properties were examined by AFM, XRD, Raman, and STEM, providing critical insights to guide further optimization of diamond GaN pn heterojunction devices.

cond-mat.mtrl-sci↗

arXiv2025

Multi-Task Surrogate-Assisted Search with Bayesian Competitive Knowledge Transfer for Expensive Optimization

Expensive optimization problems (EOPs) present significant challenges for traditional evolutionary optimization due to their limited evaluation calls. Although surrogate-assisted search (SAS) has become a popular paradigm for addressing EOPs, it still suffers from the cold-start issue. In response to this challenge, knowledge transfer has been gaining popularity for its ability to leverage search experience from potentially related instances, ultimately facilitating head-start optimization for more efficient decision-making. However, the curse of negative transfer persists when applying knowledge transfer to EOPs, primarily due to the inherent limitations of existing methods in assessing knowledge transferability. On the one hand, a priori transferability assessment criteria are intrinsically inaccurate due to their imprecise understandings. On the other hand, a posteriori methods often necessitate sufficient observations to make correct inferences, rendering them inefficient when applied to EOPs. Considering the above, this paper introduces a Bayesian competitive knowledge transfer (BCKT) method developed to improve multi-task SAS (MSAS) when addressing multiple EOPs simultaneously. Specifically, the transferability of knowledge is estimated from a Bayesian perspective that accommodates both prior beliefs and empirical evidence, enabling accurate competition between inner-task and inter-task solutions, ultimately leading to the adaptive use of promising solutions while effectively suppressing inferior ones. The effectiveness of our method in boosting various SAS algorithms for both multi-task and many-task problems is empirically validated, complemented by comparative studies that demonstrate its superiority over peer algorithms and its applicability to real-world scenarios. The source code of our method is available at https://github.com/XmingHsueh/MSAS-BCKT.

cs.NE↗

arXiv2025

R-Horizon: How Far Can Your Large Reasoning Model Really Go in Breadth and Depth?

Recent trends in test-time scaling for reasoning models (e.g., OpenAI o1, DeepSeek-R1) have led to remarkable improvements through long Chain-of-Thought (CoT). However, existing benchmarks mainly focus on immediate, single-horizon tasks, failing to adequately evaluate models' ability to understand and respond to complex, long-horizon scenarios. To address this incomplete evaluation of Large Reasoning Models (LRMs), we propose R-HORIZON, a method designed to stimulate long-horizon reasoning behaviors in LRMs through query composition. Based on R-HORIZON, we construct a long-horizon reasoning benchmark, comprising complex multi-step reasoning tasks with interdependent problems that span long reasoning horizons. Through comprehensive evaluation of LRMs using the R-HORIZON benchmark, we find that even the most advanced LRMs suffer significant performance degradation. Our analysis reveals that LRMs exhibit limited effective reasoning length and struggle to allocate thinking budget across multiple problems appropriately. Recognizing these limitations, we use R-HORIZON to construct long-horizon reasoning data for reinforcement learning with verified rewards (RLVR). Compared to training with single-horizon data, RLVR with R-HORIZON not only substantially improves performance on the multi-horizon reasoning tasks, but also promotes accuracy on standard reasoning tasks, with an increase of 7.5 on AIME2024. These results position R-HORIZON as a scalable, controllable, and low-cost paradigm for enhancing and evaluating the long-horizon reasoning capabilities of LRMs.

cs.AI↗

arXiv2025

VisualLens: Personalization through Task-Agnostic Visual History

Existing recommendation systems either rely on user interaction logs, such as online shopping history for shopping recommendations, or focus on text signals. However, item-based histories are not always accessible, and are not generalizable for multimodal recommendation. We hypothesize that a user's visual history -- comprising images from daily life -- can offer rich, task-agnostic insights into their interests and preferences, and thus be leveraged for effective personalization. To this end, we propose VisualLens, a novel framework that leverages multimodal large language models (MLLMs) to enable personalization using task-agnostic visual history. VisualLens extracts, filters, and refines a spectrum user profile from the visual history to support personalized recommendation. We created two new benchmarks, Google-Review-V and Yelp-V, with task-agnostic visual histories, and show that VisualLens improves over state-of-the-art item-based multimodal recommendations by 5-10% on Hit@3, and outperforms GPT-4o by 2-5%. Further analysis shows that VisualLens is robust across varying history lengths and excels at adapting to both longer histories and unseen content categories.

cs.CV↗

arXiv2025

VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use

Reinforcement Learning with Verifiable Rewards (RLVR) has demonstrated success in enhancing LLM reasoning capabilities, but remains limited to single-turn interactions without tool integration. While recent Agentic Reinforcement Learning with Tool use (ARLT) approaches have emerged to address multi-turn tool interactions, existing works develop task-specific codebases that suffer from fragmentation, synchronous execution bottlenecks, and limited extensibility across domains. These inefficiencies hinder broader community adoption and algorithmic innovation. We introduce VerlTool, a unified and modular framework that addresses these limitations through systematic design principles. VerlTool provides four key contributions: (1) upstream alignment with VeRL ensuring compatibility and simplified maintenance, (2) unified tool management via standardized APIs supporting diverse modalities including code execution, search, SQL databases, and vision processing, (3) asynchronous rollout execution achieving near 2$\times$ speedup by eliminating synchronization bottlenecks, and (4) comprehensive evaluation demonstrating competitive performance across 6 ARLT domains. Our framework formalizes ARLT as multi-turn trajectories with multi-modal observation tokens (text/image/video), extending beyond single-turn RLVR paradigms. We train and evaluate models on mathematical reasoning, knowledge QA, SQL generation, visual reasoning, web search, and software engineering tasks, achieving results comparable to specialized systems while providing unified training infrastructure. The modular plugin architecture enables rapid tool integration requiring only lightweight Python definitions, significantly reducing development overhead and providing a scalable foundation for tool-augmented RL research. Our code is open-sourced at https://github.com/TIGER-AI-Lab/verl-tool.

cs.AI↗

arXiv2025

OpusAnimation: Code-Based Dynamic Chart Generation

Dynamic Chart Generation (DCG) involves producing code-rendered animated visualizations as charts. While recent advances in multi-modal large language models (MLLMs) have significantly improved their capability on static chart generation and comprehension, MLLMs' potential for handling dynamic chart generation and understanding remains underexplored. To bridge this research gap, we introduce DCG-Bench (Dynamic Chart Generation Benchmark), the first benchmark evaluating MLLM's capability on dynamic chart generation tasks from three dimensions: Simple Text-to-Chart, Detailed Text-to-Chart, and Video-to-Chart tasks. We construct DCG-8K, a high-quality DCG dataset with annotations covering instruction-code-video triplets and QA pairs for both code and video evaluation. Based on DCG-8K, we explored a two-stage training recipe, proposing Joint-Code-Visual Reward for group relative policy optimization to construct expert MLLM Qwen2.5-VL-DCG-3B for the DCG task. Our benchmarking result reveals shortcomings of existing MLLMs in the visual-to-chart task, and our model beats the best open-sourced MLLM with an average 8.31% performance gain across three tasks, and shows on par performance against proprietary models with only 3B parameters, proving the effectiveness of our training recipe. Our code and dataset will be publicly available.

cs.CV↗

arXiv2025

VideoScore2: Think before You Score in Generative Video Evaluation

Recent advances in text-to-video generation have produced increasingly realistic and diverse content, yet evaluating such videos remains a fundamental challenge due to their multi-faceted nature encompassing visual quality, semantic alignment, and physical consistency. Existing evaluators and reward models are limited to single opaque scores, lack interpretability, or provide only coarse analysis, making them insufficient for capturing the comprehensive nature of video quality assessment. We present VideoScore2, a multi-dimensional, interpretable, and human-aligned framework that explicitly evaluates visual quality, text-to-video alignment, and physical/common-sense consistency while producing detailed chain-of-thought rationales. Our model is trained on a large-scale dataset VideoFeedback2 containing 27,168 human-annotated videos with both scores and reasoning traces across three dimensions, using a two-stage pipeline of supervised fine-tuning followed by reinforcement learning with Group Relative Policy Optimization (GRPO) to enhance analytical robustness. Extensive experiments demonstrate that VideoScore2 achieves superior performance with 44.35 (+5.94) accuracy on our in-domain benchmark VideoScore-Bench-v2 and 50.37 (+4.32) average performance across four out-of-domain benchmarks (VideoGenReward-Bench, VideoPhy2, etc), while providing interpretable assessments that bridge the gap between evaluation and controllable generation through effective reward modeling for Best-of-N sampling. Project Page: https://tiger-ai-lab.github.io/VideoScore2/

cs.CV↗

arXiv2025

Distinct orbital contributions to electronic and magnetic structures in La$_{4}$Ni$_{3}$O$_{10}$

High-T$_c$ superconductivity has recently been discovered in Ruddlesden-Popper phase nickelates under pressure, where the low-energy electronic structure is dominated by Ni $d_{x^2 - y^2}$ and $d_{z^2}$ orbitals. However, the respective roles of these orbitals in superconductivity remain unclear. Here, by combining X-ray absorption, electron energy loss spectroscopy, and density functional theory calculations on La$_{4}$Ni$_{3}$O$_{10}$ single crystals, we identify ligand holes in the $p_{x,y}$ orbitals of planar oxygen and the $p_z$ orbitals of apical oxygen, which hybridize with the Ni $d_{x^2-y^2}$ and $d_{z^2}$ orbitals, respectively. These ligand holes enable orbital-selective O K-edge resonant inelastic X-ray scattering (RIXS) study, which reveals that $d_{x^2-y^2}$ states dominate the low-energy charge excitations and are more itinerant. We also observe a $\sim$0.1 eV bimagnon through RIXS and Raman spectroscopy, which leads to an interlayer superexchange interaction J$_z$ of $\sim$50 meV. Our results reveal distinct contributions of Ni $d_{x^2-y^2}$ and $d_{z^2}$ orbitals to the electronic and magnetic structure and provide direct experimental insights to understand the RP-phase nickelate superconductors.

cond-mat.supr-con↗

arXiv2025

Why Bonds Fail Differently? Explainable Multimodal Learning for Multi-Class Default Prediction

In recent years, China's bond market has seen a surge in defaults amid regulatory reforms and macroeconomic volatility. Traditional machine learning models struggle to capture financial data's irregularity and temporal dependencies, while most deep learning models lack interpretability-critical for financial decision-making. To tackle these issues, we propose EMDLOT (Explainable Multimodal Deep Learning for Time-series), a novel framework for multi-class bond default prediction. EMDLOT integrates numerical time-series (financial/macroeconomic indicators) and unstructured textual data (bond prospectuses), uses Time-Aware LSTM to handle irregular sequences, and adopts soft clustering and multi-level attention to boost interpretability. Experiments on 1994 Chinese firms (2015-2024) show EMDLOT outperforms traditional (e.g., XGBoost) and deep learning (e.g., LSTM) benchmarks in recall, F1-score, and mAP, especially in identifying default/extended firms. Ablation studies validate each component's value, and attention analyses reveal economically intuitive default drivers. This work provides a practical tool and a trustworthy framework for transparent financial risk modeling.

q-fin.RM↗