Searcharxiv⌕ Search

arXiv subjects

Yan Gao

Publications and source records attributed to Yan Gao.

At least 55 records · Page 3Linked to original sources
arXiv2026

Knowledge-Graph Paths as Intermediate Supervision for Self-Evolving Search Agents

Self-evolving search agents reduce reliance on human-written training questions by generating and solving their own search tasks. We build on Search Self-Play (SSP), a representative Proposer and Solver framework in which questions are generated and answered via multi-step search and reasoning. In practice, however, SSP faces two bottlenecks: the Proposer constructs questions from isolated answer entities without relational context, yielding many invalid or unverifiable questions in early self-play training, while the Solver receives only a binary outcome reward that discards useful signal from partially on-track search trajectories. We address both bottlenecks by reusing knowledge-graph paths as construction-derived intermediate supervision for both question construction and reward shaping. First, we ground question construction in LLM-guided knowledge-graph subgraphs, providing relational context for the Proposer. Second, we observe that constructing and solving a multi-hop question can involve overlapping intermediate entities: the factual bridges used to formulate the question may provide approximate waypoints for answering it. Exploiting this overlap, we introduce Waypoint Coverage Reward (WCR), which grants graded partial credit to incorrect Solver trajectories according to their coverage of entities on the construction path, while preserving full reward for correct answers. Across seven QA benchmarks and nine model configurations, our approach improves the average score over standard SSP in all configurations, including notable gains on multi-hop QA tasks. These results suggest that knowledge-graph paths can be reused as lightweight intermediate supervision, providing both relational guidance and process feedback without additional task-specific human annotations or manually labeled process steps.

cs.AI↗
arXiv2026

Multi-modal Relational Item Representation Learning for Inferring Substitutable and Complementary Items

We study the problem of inferring substitutable and complementary items, which underpins applications such as alternative and follow-up purchase suggestions. Existing approaches typically learn from behavior-derived item-item associations using GNNs or leverage item content alone. However, these methods often overlook two key challenges: (i) user behaviors (e.g., co-view/co-purchase) only provide noisy weak supervision, and (ii) behavior signals are long-tailed, leaving many items with sparse associations. We propose MMSC, a self-supervised multi-modal relational representation learning framework that combines a multi-modal foundation model adapted to encode item metadata and a self-supervised denoising module that learns relationship-aware representations from noisy user behaviors, unified by a hierarchical aggregation mechanism. We further use LLM-assisted supervision to mitigate noise in behavior-derived supervision during training. Experiments on five real-world datasets show that MMSC consistently outperforms existing baselines by 26.1% for substitutable and 39.2% for complementary item inference, while remaining effective for cold-start items. We share our code for reproducibility.

cs.IR↗
arXiv2026

Focus on the Core: Empowering Diffusion Large Language Models by Self-Contrast

The iterative denoising paradigm of Diffusion Large Language Models (DLMs) endows them with a distinct advantage in global context modeling. However, current decoding strategies fail to leverage this capability, typically exhibiting a local preference that overlooks the heterogeneous information density within the context, ultimately degrading generation quality. To address this limitation, we systematically investigate high-information-density (HD) tokens and present two key findings: (1) explicitly conditioning on HD tokens substantially improves output quality; and (2) HD tokens exhibit an early-decoding tendency, converging earlier than surrounding tokens. Motivated by these findings, we propose Focus on the Core \textbf{(FoCore)}, a training-free decoding strategy that utilizes HD tokens in a self-contrast manner, wherein HD tokens are temporarily remasked as negative samples, to guide generation. We further introduce FoCore\_Accelerate \textbf{(FoCore\_A)}, an efficient variant that, upon detecting HD token convergence, performs parallel decoding over stable candidates within a local context window, substantially accelerating generation. Extensive experiments on math, code and logical reasoning benchmarks demonstrate that FoCore consistently improves generation quality and efficiency across both LLaDA and Dream backbones. For instance, on HumanEval, FoCore improves pass@1 from 39.02 to 42.68 over standard Classifier-Free Guidance, while FoCore-A reduces the number of decoding steps by 2.07x and per-sample latency from 20.76s to 8.64s (-58.4\%).

cs.CL↗
arXiv2026

R3A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content Platforms

Retrieval-augmented generation (RAG) plays a critical role in user-generated content (UGC) platforms, but its effectiveness critically depends on accurate query-document relevance assessment. Despite recent advances in applying large language models (LLMs) to relevance modeling, UGC platforms present unique challenges: 1) ambiguous user intent due to sparse user feedback in RAG scenarios, and 2) asymmetric relevance, where relevance is driven by localized answer-bearing content rather than global query-document similarity. To address these issues, we propose the Reinforced Reasoning model for Relevance Assessment (R3A), which decomposes relevance assessment into intent inference and evidence grounding. R3A leverages auxiliary high-clicked documents to infer latent query intent, and extracts verbatim evidence fragments to ground relevance decisions, reducing noise sensitivity and improving asymmetric relevance modeling. Experimental results demonstrate that R3A substantially outperforms strong baselines on offline benchmarks, while the distilled R3A-1.5B model achieves significant gains in large-scale online A/B testing, effectively balancing performance and practical deployability.

cs.IR↗
arXiv2026

Finch: Benchmarking Finance & Accounting across Spreadsheet-Centric Enterprise Workflows

We introduce FinWorkBench (a.k.a. Finch) for evaluating AI agents on real-world, enterprise-grade finance and accounting workflows that interleave data entry, structuring, formatting, web search, cross-file retrieval, calculation, modeling, validation, translation, visualization, and reporting. Finch is sourced from authentic enterprise workspaces from Enron (15,000 files and 500,000 emails) and other financial institutions, covering the period 2000--2025 and preserving the in-the-wild messiness of multimodal artifacts such as tables and charts across diverse domains including budgeting, trading, asset management, and operational management. We propose a workflow construction process that combines LLM-assisted mining of workflows from authentic enterprise environments with expert annotation: (1) LLM-assisted, expert-verified derivation of workflows from real-world email threads and spreadsheet version histories, and (2) meticulous annotation requiring over 700 hours of expert effort. This yields 172 composite workflows with 384 tasks, involving 1,710 spreadsheets with 27 million cells, along with PDFs and other artifacts, capturing the intrinsically messy, long-horizon, knowledge-intensive, and collaborative nature of real-world enterprise work. We conduct both human and automated evaluations of frontier AI systems, including GPT-5.1, Claude Sonnet 4.5, Claude Opus 4.5, Gemini 3 Pro, Grok 4, and Qwen 3 Max. Under human evaluation, GPT-5.1 Pro spends an average of 16.8 minutes per workflow yet passes only 38.4% of workflows. Comprehensive case studies further surface the challenges that real-world enterprise workflows pose for AI agents.

cs.AI↗
arXiv2026

Symmetry-protected coexistence of a nodal surface and multiple types of Weyl fermions in $P6_3$-$\text{B}_{30}$

The coexistence of topological states with different dimensionalities in a single crystalline system offers a unique platform to study the interplay of distinct fermionic excitations. Here, integrating first-principles calculations with symmetry analysis, we propose the three-dimensional boron allotrope $P6_3$-$\text{B}_{30}$ as an ideal, structurally stable candidate for exploring multidimensional topological physics. Benefiting from the practically negligible spin-orbit coupling of the light-element framework, $P6_3$-$\text{B}_{30}$ operates as a pristine spinless topological semimetal. We show that the combined time-reversal and twofold screw symmetry ($\mathcal{T}S_{2z}$) enforces a robust two-dimensional nodal surface on the $k_z = π$ plane via a Kramers-like degeneracy. Concurrently, the system hosts a diverse set of zero-dimensional Weyl fermions -- including an unconventional double-Weyl point ($\mathcal{C} = -2$), conventional Type-I WPs ($\mathcal{C} = -1$), and completely tilted Type-II WPs ($\mathcal{C} = +1$) -- emerging at the high-symmetry points $Γ$ and K, as well as along the H-K path, protected by $C_6$ and $C_3$ crystalline rotational symmetries. Crucially, the substantial momentum-space separation between the nodal surface and Weyl points allows for their unambiguous independent resolution. Calculations of the (100) surface states reveal distinct, nontrivial Fermi arcs connecting Weyl nodes of opposite chirality. This work establishes $P6_3$-$\text{B}_{30}$ as a compelling material platform for investigating the physics of multidimensional hybrid topological fermions and their interplay.

cond-mat.mtrl-sci↗
arXiv2026

SPARD: Self-Paced Curriculum for RL Alignment via Integrating Reward Dynamics and Data Utility

The evolution of Large Language Models (LLMs) is shifting the focus from single, verifiable tasks toward complex, open-ended real-world scenarios, imposing significant challenges on the post-training phase. In these settings, the scale and complexity of reward systems have grown significantly, transitioning toward multi-objective formulations that encompass a comprehensive spectrum of model capabilities and application contexts. However, traditional methods typically rely on fixed reward weights, ignoring non-stationary learning dynamics and struggling with data heterogeneity across dimensions. To address these issues, we propose SPARD, a framework that establishes an automated, self-paced curriculum by perceiving learning progress to dynamically adjust multi-objective reward weights and data importance, thereby synchronizing learning intent with data utility for optimal performance. Extensive experiments across multiple benchmarks demonstrate that SPARD significantly enhances model capabilities across all domains.

cs.AI↗
arXiv2026

Logics-Parsing-Omni Technical Report

Addressing the challenges of fragmented task definitions and the heterogeneity of unstructured data in multimodal parsing, this paper proposes the Omni Parsing framework. This framework establishes a Unified Taxonomy covering documents, images, and audio-visual streams, introducing a progressive parsing paradigm that bridges perception and cognition. Specifically, the framework integrates three hierarchical levels: 1) Holistic Detection, which achieves precise spatial-temporal grounding of objects or events to establish a geometric baseline for perception; 2) Fine-grained Recognition, which performs symbolization (e.g., OCR/ASR) and attribute extraction on localized objects to complete structured entity parsing; and 3) Multi-level Interpreting, which constructs a reasoning chain from local semantics to global logic. A pivotal advantage of this framework is its evidence anchoring mechanism, which enforces a strict alignment between high-level semantic descriptions and low-level facts. This enables ``evidence-based'' logical induction, transforming unstructured signals into standardized knowledge that is locatable, enumerable, and traceable. Building on this foundation, we constructed a standardized dataset and released the Logics-Parsing-Omni model, which successfully converts complex audio-visual signals into machine-readable structured knowledge. Experiments demonstrate that fine-grained perception and high-level cognition are synergistic, effectively enhancing model reliability. Furthermore, to quantitatively evaluate these capabilities, we introduce OmniParsingBench. Code, models and the benchmark are released at https://github.com/alibaba/Logics-Parsing/tree/master/Logics-Parsing-Omni.

cs.AI↗
arXiv2026

Symmetry-protected four double-Weyl fermions and their topological phase transitions in nonmagnetic crystals

Realizing Weyl semimetals (WSMs) with the minimal number of Weyl points (WPs) fundamentally simplifies extracting intrinsic topological responses. While a minimum of four conventional ($|C|=1$) WPs in nonmagnetic crystals is well-established, the exact symmetry requirements and material realization for the unique configuration of four unconventional double-Weyl points (DWPs, $|C|=2$) remain unresolved. Here, we establish rigorous crystalline symmetry constraints restricting the existence of exactly four symmetry-protected DWPs to merely 28 space groups in both nonmagnetic spinless and spinful systems. Guided by this classification, we identify an $sp$$^2$--$sp$$^3$ hybridized chiral carbon allotrope, THRLN-C$_{32}$, as an ideal candidate hosting precisely this four-DWP configuration near the Fermi level. These $C_4$-protected DWPs project extended or closed-loop Fermi arcs onto the surface Brillouin zone, providing unambiguous spectroscopic signatures. Furthermore, external strain drives profound topological phase transitions encapsulated in a unified evolution landscape: the pristine four-DWP state dissociates into two exotic three-terminal Weyl complexes, degenerates into eight conventional $|C|=1$ WPs, or collapses into a trivial insulator. This work provides a definitive theoretical framework for minimal double-WSMs in nonmagnetic spinful systems and introduces an optimal material platform for investigating strain-tunable topological quantum phenomena.

cond-mat.mtrl-sci↗
arXiv2026

PRCCF: A Persona-guided Retrieval and Causal-aware Cognitive Filtering Framework for Emotional Support Conversation

Emotional Support Conversation (ESC) aims to alleviate individual emotional distress by generating empathetic responses. However, existing methods face challenges in effectively supporting deep contextual understanding. To address this issue, we propose PRCCF, a Persona-guided Retrieval and Causality-aware Cognitive Filtering framework. Specifically, the framework incorporates a persona-guided retrieval mechanism that jointly models semantic compatibility and persona alignment to enhance response generation. Furthermore, it employs a causality-aware cognitive filtering module to prioritize causally relevant external knowledge, thereby improving contextual cognitive understanding for emotional reasoning. Extensive experiments on the ESConv dataset demonstrate that PRCCF outperforms state-of-the-art baselines on both automatic metrics and human evaluations. Our code is publicly available at: https://github.com/YancyLyx/PRCCF.

cs.CL↗
arXiv2026

Adaptive Federated Fine-Tuning of Self-Supervised Speech Representations

Integrating Federated Learning (FL) with self-supervised learning (SSL) enables privacy-preserving fine-tuning for speech tasks. However, federated environments exhibit significant heterogeneity: clients differ in computational capacity, causing straggler effects under unified fine-tuning, while diverse downstream tasks require different representation depths, making full-model updates inefficient. To address these challenges, we propose an adaptive federated fine-tuning framework with early exits. Lightweight prediction heads are inserted at intermediate layers of the SSL backbone, allowing clients to terminate computation based on local constraints and task requirements. We further introduce a layer-wise, depth-aware partial aggregation strategy to better utilize representations from different network depths. Experiments show that the framework reduces edge overhead, supports heterogeneous hardware, and maintains competitive performance in resource-constrained federated environments.

eess.AS↗
arXiv2026

Single-pair charge-2 Weyl-Dirac composite semimetals

The Nielsen--Ninomiya theorem requires that the total topological chiral charges in a crystal vanish, a constraint typically satisfied by identical nodes like Weyl--Weyl pairs. Whether a minimal heterogeneous configuration -- comprising a single Weyl point (WP) and a single Dirac point (DP) -- can exist in an electronic system has remained unresolved. Here, by systematically classifying all 1651 magnetic space groups (MSGs), we reveal that only 14 MSGs without spin-orbit coupling (SOC) and 10 MSGs with SOC are compatible with this exotic state. Furthermore, for nonmagnetic crystals, this configuration is uniquely realized in the spinless limit of chiral space groups 92 and 96. Guided by this principle, we predict an ideal realization in chiral three-dimensional boron allotropes (SDHBN-B$_{28}$ enantiomers). First-principles calculations unveil a $|C|=2$ WP at the $Γ$ point and a $|C|=2$ DP at the $A$ point, which constitute the only fermions near the Fermi level within a large $2$ eV energy window. Strikingly, the structural chirality rigidly dictates the sign of the topological charges, yielding two ultralong Fermi arcs spanning the surface Brillouin zone. Our work provides a complete crystallographic classification and a definitive material platform for exploring minimal heterogeneous chiral fermions.

cond-mat.mtrl-sci↗
arXiv2026

HoRD: Robust Humanoid Control via History-Conditioned Reinforcement Learning and Online Distillation

Humanoid robots can suffer significant performance drops under small changes in dynamics, task specifications, or environment setup. We propose HoRD, a two-stage learning framework for robust humanoid control under domain shift. First, we train a high-performance teacher policy via history-conditioned reinforcement learning, where the policy infers latent dynamics context from recent state--action trajectories to adapt online to diverse randomized dynamics. Second, we perform online distillation to transfer the teacher's robust control capabilities into a transformer-based student policy that operates on sparse root-relative 3D joint keypoint trajectories. By combining history-conditioned adaptation with online distillation, HoRD enables a single policy to adapt zero-shot to unseen domains without per-domain retraining. Extensive experiments show HoRD outperforms strong baselines in robustness and transfer, especially under unseen domains and external perturbations. Code and project page are available at https://tonywang-0517.github.io/hord/.

cs.RO↗
arXiv2026

Edge-Cloud Collaborative Speech Emotion Captioning via Token-Level Speculative Decoding in Audio-Language Models

Speech Emotion Captioning (SEC) leverages large audio-language models to generate rich, context-aware affective descriptions from speech. However, real-world deployment remains challenging due to the substantial computational demands on resource-constrained edge devices and the privacy risks of transmitting biometric audio. While smaller audio-language models enable efficient on-device SEC, their limited capacity often weakens subtle paralinguistic modeling and fine-grained affective grounding. We propose an edge-cloud collaborative framework based on Uncertainty-Guided Speculative Decoding (UGSD). A lightweight edge model drafts captions locally, and only high-uncertainty token blocks are selectively escalated to a stronger cloud verifier for validation. Experiments on the MER2024 benchmark demonstrate substantial BLEU improvements up to 62.7%. UGSD further achieves 1.4x lower latency and 8.5x higher token throughput compared to an edge-only model. These results empirically characterize the quality-efficiency-privacy trade-off in deployable SEC systems.

cs.SD↗
arXiv2026

Aligning Large Language Models with Searcher Preferences

The paradigm shift from item-centric ranking to answer-centric synthesis is redefining the role of search engines. While recent industrial progress has applied generative techniques to closed-set item ranking in e-commerce, research and deployment of open-ended generative search on large content platforms remain limited. This setting introduces challenges, including robustness to noisy retrieval, non-negotiable safety guarantees, and alignment with diverse user needs. In this work, we introduce SearchLLM, the first large language model (LLM) for open-ended generative search. We design a hierarchical, multi-dimensional reward system that separates bottom-line constraints, including factual grounding, basic answer quality and format compliance, from behavior optimization objectives that promote robustness to noisy retrieval and alignment with user needs. Concretely, our reward model evaluates responses conditioned on the user query, session history, and retrieved evidence set, combining rule-based checks with human-calibrated LLM judges to produce an interpretable score vector over these dimensions. We introduce a Gated Aggregation Strategy to derive the training reward for optimizing SearchLLM with Group Relative Policy Optimization (GRPO). We deploy SearchLLM in the AI search entry of RedNote. Offline evaluations and online A/B tests show improved generation quality and user engagement, increasing Valid Consumption Rate by 1.03% and reducing Re-search Rate by 2.81%, while upholding strict safety and reliability standards.

cs.CL↗
arXiv2026

Can Vision Language Models Assess Graphic Design Aesthetics? A Benchmark, Evaluation, and Dataset Perspective

Assessing the aesthetic quality of graphic design is central to visual communication, yet remains underexplored in vision language models (VLMs). We investigate whether VLMs can evaluate design aesthetics in ways comparable to humans. Prior work faces three key limitations: benchmarks restricted to narrow principles and coarse evaluation protocols, a lack of systematic VLM comparisons, and limited training data for model improvement. In this work, we introduce AesEval-Bench, a comprehensive benchmark spanning four dimensions, twelve indicators, and three fully quantifiable tasks: aesthetic judgment, region selection, and precise localization. Then, we systematically evaluate proprietary, open-source, and reasoning-augmented VLMs, revealing clear performance gaps against the nuanced demands of aesthetic assessment. Moreover, we construct a training dataset to fine-tune VLMs for this domain, leveraging human-guided VLM labeling to produce task labels at scale and indicator-grounded reasoning to tie abstract indicators to concrete design regions.Together, our work establishes the first systematic framework for aesthetic quality assessment in graphic design. Our code and dataset will be released at: \href{https://github.com/arctanxarc/AesEval-Bench}{https://github.com/arctanxarc/AesEval-Bench}

cs.CV↗
arXiv2026

Pix2Key: Controllable Open-Vocabulary Retrieval with Semantic Decomposition and Self-Supervised Visual Dictionary Learning

Composed Image Retrieval (CIR) uses a reference image plus a natural-language edit to retrieve images that apply the requested change while preserving other relevant visual content. Classic fusion pipelines typically rely on supervised triplets and can lose fine-grained cues, while recent zero-shot approaches often caption the reference image and merge the caption with the edit, which may miss implicit user intent and return repetitive results. We present Pix2Key, which represents both queries and candidates as open-vocabulary visual dictionaries, enabling intent-aware constraint matching and diversity-aware reranking in a unified embedding space. A self-supervised pretraining component, V-Dict-AE, further improves the dictionary representation using only images, strengthening fine-grained attribute understanding without CIR-specific supervision. On the DFMM-Compose benchmark, Pix2Key improves Recall@10 up to 3.2 points, and adding V-Dict-AE yields an additional 2.3-point gain while improving intent consistency and maintaining high list diversity.

cs.CV↗
arXiv2026

Symmetry-Protected Minimum of Four Conventional Weyl Points in Nonmagnetic Crystals

Realizing nonmagnetic Weyl semimetals (WSMs) with the minimal number of conventional Weyl points (WPs) and a clean Fermi surface remains a central challenge. Here, combining symmetry analysis with first-principles calculations, we establish the definitive conditions under which a nonmagnetic crystal can host exactly four conventional ($C = \pm 1$) WPs, identifying 76 space groups in the spinless limit and 83 in the spinful case that allow this minimal configuration. Guided by this framework, we predict two previously unknown boron allotropes, P6-B$_{48}$ and TBIN-B$_{48}$, as ideal WSMs. Both exhibits precisely four isolated WPs near the Fermi level, with exceptionally clean electronic structures. Notably, the WPs in P6-B$_{48}$ are pinned to high-symmetry points, while those in TBIN-B$_{48}$ lie along high-symmetry lines, leading to distinct and experimentally accessible surface states, including single and double Fermi arcs. Our work provides a complete symmetry-based foundation and pristine material platforms for minimal Weyl physics.

cond-mat.mtrl-sci↗