arXiv subjects

Jie Yang

Publications and source records attributed to Jie Yang.

At least 109 records · Page 6Linked to original sources

arXiv2026

A Survey of Reinforcement Learning for Large Language Models under Data Scarcity: Challenges and Solutions

Reinforcement learning (RL) has emerged as a powerful post-training paradigm for enhancing the reasoning capabilities of large language models (LLMs). However, reinforcement learning for LLMs faces substantial data scarcity challenges, including the limited availability of high-quality external supervision and the constrained volume of model-generated experience. These limitations make data-efficient reinforcement learning a critical research direction. In this survey, we present the first systematic review of reinforcement learning for LLMs under data scarcity. We propose a bottom-up hierarchical framework built around three complementary perspectives: the data-centric perspective, the training-centric perspective, and the framework-centric perspective. We develop a taxonomy of existing methods, summarize representative approaches in each category, and analyze their strengths and limitations. Our taxonomy aims to provide a clear conceptual foundation for understanding the design space of data-efficient RL for LLMs and to guide researchers working in this emerging area. We hope this survey offers a comprehensive roadmap for future research and inspires new directions toward more efficient and scalable reinforcement learning post-training for LLMs.

cs.LG↗

arXiv2026

FinTruthQA: A Benchmark for AI-Driven Financial Disclosure Quality Assessment in Investor -- Firm Interactions

Accurate and transparent financial information disclosure is essential for market efficiency, investor decision-making, and corporate governance. Chinese stock exchanges' investor interactive platforms provide a widely used channel through which listed firms respond to investor concerns, yet these responses are often limited or non-substantive, making disclosure quality difficult to assess at scale. To address this challenge, we introduce FinTruthQA, to our knowledge the first benchmark for AI-driven assessment of financial disclosure quality in investor-firm interactions. FinTruthQA comprises 6,000 real-world financial Q&A entries, each manually annotated based on four key evaluation criteria: question identification, question relevance, answer readability, and answer relevance. We benchmark statistical machine learning models, pre-trained language models and their fine-tuned variants, as well as large language models (LLMs), on FinTruthQA. Experiments show that existing models achieve strong performance on question identification and question relevance (F1 > 95%), but remain substantially weaker on answer readability (Micro F1 approximately 88%) and especially answer relevance (Micro F1 approximately 80%), highlighting the nontrivial difficulty of fine-grained disclosure quality assessment. Domain- and task-adapted pre-trained language models consistently outperform general-purpose models and LLM-based prompting on the most challenging settings. These findings position FinTruthQA as a practical foundation for AI-driven disclosure monitoring in capital markets, with value for regulatory oversight, investor protection, and disclosure governance in real-world financial settings.

cs.CL↗

arXiv2026

MinerU2.5-Pro: Pushing the Limits of Data-Centric Document Parsing at Scale

Current document parsing methods advance primarily through model architecture innovation, while systematic engineering of training data remains underexplored. Yet state-of-the-art models spanning diverse architectures and parameter scales exhibit highly consistent failure patterns on the same set of hard samples, suggesting that the performance bottleneck stems from shared deficiencies in training data rather than from architectural differences. Building on this finding, we present MinerU2.5-Pro, which advances the state of the art purely through data engineering and training strategy design while retaining the 1.2B-parameter architecture of MinerU2.5 unchanged. At its core is a Data Engine co-designed around coverage, informativeness, and annotation accuracy: Diversity-and-Difficulty-Aware Sampling expands training data from under 10M to 65.5M samples while mitigating distribution shift; Cross-Model Consistency Verification leverages output consensus among heterogeneous models to assess sample difficulty and generate reliable annotations; the Judge-and-Refine pipeline improves annotation quality for hard samples through render-then-verify iterative correction. A three-stage progressive training strategy--large-scale pre-training, hard sample fine-tuning, and GRPO alignment--sequentially exploits these data at different quality tiers. On the evaluation front, we rectify element-matching biases in OmniDocBench v1.5 and introduce a Hard subset, establishing the more discriminative OmniDocBench v1.6 protocol. Without any architectural modification, MinerU2.5-Pro achieves 95.69 on OmniDocBench v1.6, improving over the same-architecture baseline by 2.71 points and surpassing all existing methods, including those based on models with over 200x more parameters.

cs.CV↗

arXiv2026

Cascade of Spin Liquids in a Bilayer Triangular-lattice Antiferromagnet Rb_2Co_2(SeO_3)_3

In frustrated Ising magnets, classical spin liquids (CSLs) with macroscopic ground-state degeneracy can survive against conventional magnetic order, as exemplified by systems on triangular, kagome and pyrochlore lattices at zero field. Here we report the discovery of a high-field route toward spin liquids in a bilayer triangular lattice antiferromagnet, Rb$_2$Co$_2$(SeO$_3$)$_3$. We demonstrate that a cascade of CSLs -- characterized by doubly degenerate one-up-one-down local spin configurations and a residual entropy of 1/2(1-M/M_s)Rln2 per mole -- emerges through field-controlled dilution of Ising dimers. Owing to the interplay of intra- and inter-layer interactions, these CSLs are further stabilized by lattice symmetry breaking at fractional magnetization plateaus. Such field-induced spin liquids can be understood as a consequence of generalized ice rules, analogous to those governing in pyrochlore antiferromagnets. In particular, the 5/6-plateau state is a candidate quantum spin liquid. Our results thereby establish a new pathway for exploring diverse spin liquid states across both classical and quantum regimes.

cond-mat.str-el↗

arXiv2026

From Poisson Observations to Fitted Negative Binomial Distribution

The negative binomial distribution has been widely used as a more flexible model than the Poisson distribution for count data. However, when the true data-generating process is Poisson, it is often challenging to distinguish it from a negative binomial distribution with extreme parameter values, and existing maximum likelihood estimation procedures for the negative binomial distribution may fail or produce unstable estimates. To address this issue, we develop a new algorithm for computing the maximum likelihood estimate of negative binomial parameters, which is more efficient and more accurate than existing methods. We further extend negative binomial distributions with a new parameterization to cover Poisson distributions as a special class. We provide theoretical justifications showing that, when applied to a Poisson data, the estimated parameters of the extended negative binomial distribution can consistently recover the true Poisson distribution.

math.ST↗

arXiv2026

Expected Weighted D-optimal Designs for Experiments with Mixed Factors

Optimal designs can help experimenters obtain more accurate parameter estimates with reduced experimental time and cost. In this paper, we characterize the Expected Weighted (EW) D-optimal designs as robust designs against unknown parameter values for experiments under a general parametric model with discrete and continuous factors. When a pilot study is available, we recommend sample-based EW D-optimal designs for subsequent experiments. Otherwise, we recommend EW D-optimal designs under a prior distribution for model parameters. We propose an EW ForLion algorithm for finding EW D-optimal designs with mixed factors, and justify that the designs found by our algorithm are EW D-optimal. To facilitate potential users in practice, we also develop a rounding algorithm that converts an approximate design with mixed factors to exact designs with prespecified grid points and the total number of experimental units. By applying our algorithms for real experiments under multinomial logistic models or generalized linear models, we show that our designs are highly efficient with respect to locally D-optimal designs and more robust against parameter value misspecifications.

stat.ME↗

arXiv2026

SCP: Spatial Causal Prediction in Video

Spatial reasoning, the ability to understand spatial relations, causality, and dynamic evolution, is central to human intelligence and essential for real-world applications such as autonomous driving and robotics. Existing studies, however, primarily assess models on visible spatio-temporal understanding, overlooking their ability to infer unseen past or future spatial states. In this work, we introduce Spatial Causal Prediction (SCP), a new task paradigm that challenges models to reason beyond observation and predict spatial causal outcomes. We further construct SCP-Bench, a benchmark comprising 2,500 QA pairs across 1,181 videos spanning diverse viewpoints, scenes, and causal directions, to support systematic evaluation. Through comprehensive experiments on {23} state-of-the-art models, we reveal substantial gaps between human and model performance, limited temporal extrapolation, and weak causal grounding. We further analyze key factors influencing performance and propose perception-enhancement and reasoning-guided strategies toward advancing spatial causal intelligence. The project page is https://guangstrip.github.io/SCP-Bench.

cs.CV↗

arXiv2026

Few-Shot Distribution-Aligned Flow Matching for Data Synthesis in Medical Image Segmentation

Data heterogeneity hinders clinical deployment of medical image analysis models, and generative data augmentation helps mitigate this issue. However, recent diffusion-based methods that synthesize image-mask pairs often ignore distribution shifts between generated and real images across scenarios, and such mismatches can markedly degrade downstream performance. To address this issue, we propose AlignFlow, a flow matching model that aligns with the target reference image distribution via differentiable reward fine-tuning, and remains effective even when only a small number of reference images are provided. Specifically, we divide the training of the flow matching model into two stages: in the first stage, the model fits the training data to generate plausible images; Then, we introduce a distribution alignment mechanism and employ differentiable reward to steer the generated images toward the distribution of the given samples from the target domain. In addition, to enhance the diversity of generated masks, we also design a flow matching based mask generation to complement the diversity in regions of interest. Extensive experiments demonstrate the effectiveness of our approach, i.e., performance improvement by 3.5-4.0% in mDice and 3.5-5.6% in mIoU across a variety of datasets and scenarios.

eess.IV↗

arXiv2026

Integrating Low-Altitude SAR Imaging into UAV Data Backhaul

Synthetic aperture radar (SAR) on unmanned aerial vehicles (UAVs) enables high-resolution sensing in low-altitude wireless networks, while requiring reliable uplink data backhaul to ground base stations under dynamic channel conditions. Conventional orthogonal frequency division multiplexing (OFDM)-based SAR systems rely on pilot or deterministic signaling, which occupies only a small fraction of the available timefrequency (TF) resources and limits imaging performance. This paper develops a data-aided OFDM-SAR imaging framework that reuses uplink communication data symbols for sensing, thereby exploiting the dominant TF resources of the UAV backhaul link. However, the randomness of data symbols disrupts the coherent structure required for SAR imaging, especially in highly dynamic channels with strong TF coupling, leading to severe degradation in range-Doppler focusing. To address this issue, we establish a unified TF domain filtering framework to suppress data-induced randomness and recover an equivalent deterministic imaging channel. Within this framework, reciprocal, matched, and Wiener filtering are interpreted under a common formulation, enabling a systematic characterization of their impact on imaging performance. A normalized mean square error (NMSE) metric of a reference point target's profile is further adopted to quantify the joint effects of randomnessinduced distortion and noise amplification. Simulation results based on 5G NR parameters show that the proposed dataaided scheme significantly outperforms pilot-only approaches by leveraging uplink data resources, demonstrating that effective TF-domain filtering is essential to ensure high-resolution imaging in dynamic UAV channels.

eess.SP↗

arXiv2026

BRIDGE: Benchmarking Large Language Models for Understanding Real-world Clinical Practice Text

Large language models (LLMs) hold great promise for medical applications and are evolving rapidly, with new models being released at an accelerated pace. However, benchmarking on large-scale real-world data such as electronic health records (EHRs) is critical, as clinical decisions are directly informed by these sources, yet current evaluations remain limited. Most existing benchmarks rely on medical exam-style questions or PubMed-derived text, failing to capture the complexity of real-world clinical data. Others focus narrowly on specific application scenarios, limiting their generalizability across broader clinical use. To address this gap, we present BRIDGE, a comprehensive multilingual benchmark comprising 87 tasks sourced from real-world clinical data sources across nine languages. It covers eight major task types spanning the entire continuum of patient care across six clinical stages and 20 representative applications, including triage and referral, consultation, information extraction, diagnosis, prognosis, and billing coding, and involves 14 clinical specialties. We systematically evaluated 95 LLMs (including DeepSeek-R1, GPT-4o, Gemini series, and Qwen3 series) under various inference strategies. Our results reveal substantial performance variation across model sizes, languages, natural language processing tasks, and clinical specialties. Notably, we demonstrate that open-source LLMs can achieve performance comparable to proprietary models, while medically fine-tuned LLMs based on older architectures often underperform versus updated general-purpose models. The BRIDGE and its corresponding leaderboard serve as a foundational resource and a unique reference for the development and evaluation of new LLMs in real-world clinical text understanding. The BRIDGE leaderboard: https://huggingface.co/spaces/YLab-Open/BRIDGE-Medical-Leaderboard

cs.CL↗

arXiv2026

LongCat-Next: Lexicalizing Modalities as Discrete Tokens

The prevailing Next-Token Prediction (NTP) paradigm has driven the success of large language models through discrete autoregressive modeling. However, contemporary multimodal systems remain language-centric, often treating non-linguistic modalities as external attachments, leading to fragmented architectures and suboptimal integration. To transcend this limitation, we introduce Discrete Native Autoregressive (DiNA), a unified framework that represents multimodal information within a shared discrete space, enabling a consistent and principled autoregressive modeling across modalities. A key innovation is the Discrete Native Any-resolution Visual Transformer (dNaViT), which performs tokenization and de-tokenization at arbitrary resolutions, transforming continuous visual signals into hierarchical discrete tokens. Building on this foundation, we develop LongCat-Next, a native multimodal model that processes text, vision, and audio under a single autoregressive objective with minimal modality-specific design. As an industrial-strength foundation model, it excels at seeing, painting, and talking within a single framework, achieving strong performance across a wide range of multimodal benchmarks. In particular, LongCat-Next addresses the long-standing performance ceiling of discrete vision modeling on understanding tasks and provides a unified approach to effectively reconcile the conflict between understanding and generation. As an attempt toward native multimodality, we open-source the LongCat-Next and its tokenizers, hoping to foster further research and development in the community. GitHub: https://github.com/meituan-longcat/LongCat-Next

cs.CV↗

arXiv2026

Higher descent equations based on 2-term $L_{\infty}$ algebras

In this paper, we develop the higher descent equations for higher gauge theories within the framework of 2-term $L_{\infty}$ algebras. Starting from a multilinear symmetric invariant polynomial, we construct a family of higher Chern-Simons type characteristic classes and verify that they satisfy the higher descent equations. These polynomials encode both the higher Chern-Weil theorem and the higher gauge anomalies.

hep-th↗

arXiv2026

Accurate Precipitation Forecast by Efficiently Learning from Massive Atmospheric Variables and Unbalanced Distribution

Short-term (0-24 hours) precipitation forecasting is highly valuable to socioeconomic activities and public safety. However, the highly complex evolution patterns of precipitation events, the extreme imbalance between precipitation and non-precipitation samples, and the inability of existing models to efficiently and effectively utilize large volumes of multi-source atmospheric observation data hinder improvements in precipitation forecasting accuracy and computational efficiency. To address the above challenges, this study developed a novel forecasting model capable of effectively and efficiently utilizing massive atmospheric observations by automatically extracting and iteratively predicting the latent features strongly associated with precipitation evolution. Furthermore, this study introduces a 'WMCE' loss function, designed to accurately discriminate extremely scarce precipitation events while precisely predicting their intensity values. Extensive experiments on two datasets demonstrate that our proposed model substantially and consistently outperforms all prevalent baselines in both accuracy and efficiency. Moreover, the proposed forecasting model substantially lowers the computational cost required to obtain valuable predictions compared to existing approaches, thereby positioning it as a milestone for efficient and practical precipitation forecasting.

cs.LG↗

arXiv2026

Bridging Past and Future: Distribution-Aware Alignment for Time Series Forecasting

Although contrastive and other representation-learning methods have long been explored in vision and NLP, their adoption in modern time series forecasters remains limited. We believe they hold strong promise for this domain. To unlock this potential, we explicitly align past and future representations, thereby bridging the distributional gap between input histories and future targets. To this end, we introduce TimeAlign, a lightweight, plug-and-play framework that establishes a new representation paradigm, distinct from contrastive learning, by aligning auxiliary features via a simple reconstruction task and feeding them back into any base forecaster. Extensive experiments across eight benchmarks verify its superior performance. Further studies indicate that the gains arise primarily from correcting frequency mismatches between historical inputs and future outputs. Additionally, we provide two theoretical justifications for how reconstruction improves forecasting generalization and how alignment increases the mutual information between learned representations and predicted targets. The code is available at https://github.com/TROUBADOUR000/TimeAlign.

cs.LG↗

arXiv2026

Shape Polarization and Quasiparticle Alignment in the [523]5/2 and [642]5/2 bands of $^{169}$Hf

Rotational properties of [523]5/2 (h11/2) and [642]5/2 (i13/2) signature partner bands in 169Hf are investigated using the Total Routhian Surface (TRS) method. Experimental data show a distinct signature inversion in the [523]5/2 band at high spin, while the [642]5/2 band exhibits conventional signature splitting. Our analysis identifies a proton subshell gap at Z=72 (beta2 ~ 0.35) that 'locks' the proton core, allowing neutron-driven dynamics to dominate signature staggering. A critical shape bifurcation is identified in the [523]5/2 configuration: the alpha = -1/2 branch remains rigid at high deformation (beta2 ~ 0.32, gamma ~ -10 deg), whereas the alpha = +1/2 branch transitions toward reduced beta2 ~ 0.20 and enhanced hexadecapole deformation. This shift facilitates prompt i13/2 neutron alignment at h-bar omega ~ 0.3 MeV for the alpha = +1/2 branch. Furthermore, the alpha = -1/2 signature enters a gamma-soft regime at high spin (gamma fluctuating between -10 and +10 deg), favoring the unfavored signature and reinforcing the Routhian crossing. In contrast, the [642]5/2 band maintains a stable triaxial shape (gamma ~ -18 deg), preventing inversion. At h-bar omega ~ 0.5 MeV, a "shape-jump" to beta2 ~ 0.38 is predicted for the i13/2 band, signaling the breakthrough of the Z=72 gap and the onset of a highly deformed proton-aligned regime.

nucl-th↗

arXiv2026

No Need For Real Anomaly: MLLM Empowered Zero-Shot Video Anomaly Detection

The collection and detection of video anomaly data has long been a challenging problem due to its rare occurrence and spatio-temporal scarcity. Existing video anomaly detection (VAD) methods under perform in open-world scenarios. Key contributing factors include limited dataset diversity, and inadequate understanding of context-dependent anomalous semantics. To address these issues, i) we propose LAVIDA, an end-to-end zero-shot video anomaly detection framework. ii) LAVIDA employs an Anomaly Exposure Sampler that transforms segmented objects into pseudo-anomalies to enhance model adaptability to unseen anomaly categories. It further integrates a Multimodal Large Language Model (MLLM) to bolster semantic comprehension capabilities. Additionally, iii) we design a token compression approach based on reverse attention to handle the spatio-temporal scarcity of anomalous patterns and decrease computational cost. The training process is conducted solely on pseudo anomalies without any VAD data. Evaluations across four benchmark VAD datasets demonstrate that LAVIDA achieves SOTA performance in both frame-level and pixel-level anomaly detection under the zero-shot setting. Our code is available in https://github.com/VitaminCreed/LAVIDA.

cs.CV↗

arXiv2026

Beam Prediction Based on Multimodal Large Language Models

Accurate beam prediction is a key enabler for next-generation wireless communication systems. In this paper, we propose a multimodal large language model (LLM)-based beam prediction framework that effectively utilizes contextual information, provided by sensory data including RGB camera images and LiDAR point clouds. To effectively fuse heterogeneous modalities, we design specialized modality encoders together with a beam-guided attention masking mechanism and a high-frequency temporal alignment strategy, enabling robust cross-modal feature integration under dynamic environments. Furthermore, we construct a large-scale multimodal dataset for communication, named Multimodal-Wireless, which covers diverse weather and traffic conditions with high-fidelity ray-tracing labels. Extensive simulation results demonstrate that the proposed approach significantly reduces the reliance on oracle angle-of-departure knowledge and consistently outperforms state-of-the-art multimodal LLM-based beam prediction methods in terms of beam accuracy and communication performance, improving the average Top-1 accuracy to 80.8% and the average normalized gain to 89.1%.

eess.SP↗

arXiv2026

A-optimal Designs under Generalized Linear Models

Designing efficient experiments under practical constraints is critical in both scientific research and industrial practice. Focusing on minimizing the average variance of the parameter estimates, A-optimal designs show advantages in screening factors and reducing prediction errors. Compared with other criteria, however, algorithms and software for generating A-optimal designs are scarce. In this paper, we characterize A-optimal designs under generalized linear models theoretically and develop efficient algorithms for identifying them. When a predetermined finite set of experimental settings is given, we derive analytic solutions or establish necessary and sufficient conditions for obtaining A-optimal approximate allocations. We show that a lift-one algorithm based on our formulae outperforms commonly used algorithms for finding A-optimal allocations. When continuous factors or design regions get involved, we develop a ForLion algorithm that is guaranteed to find A-optimal designs with mixed factors. Numerical studies show that our algorithms can find highly efficient designs with reduced numbers of distinct experimental settings, which may save both experimental time and cost significantly. Along with a rounding-off algorithm that converts approximate allocations to exact ones, we demonstrate that stratified samplers based on A-optimal allocations may provide more accurate parameter estimates than commonly used samplers.

stat.ME↗