arXiv subjects

Yue Wang

Publications and source records attributed to Yue Wang.

At least 361 records · Page 20Linked to original sources

arXiv2025

Sentient Agent as a Judge: Evaluating Higher-Order Social Cognition in Large Language Models

Assessing how well a large language model (LLM) understands human, rather than merely text, remains an open challenge. To bridge the gap, we introduce Sentient Agent as a Judge (SAGE), an automated evaluation framework that measures an LLM's higher-order social cognition. SAGE instantiates a Sentient Agent that simulates human-like emotional changes and inner thoughts during interaction, providing a more realistic evaluation of the tested model in multi-turn conversations. At every turn, the agent reasons about (i) how its emotion changes, (ii) how it feels, and (iii) how it should reply, yielding a numerical emotion trajectory and interpretable inner thoughts. Experiments on 100 supportive-dialogue scenarios show that the final Sentient emotion score correlates strongly with Barrett-Lennard Relationship Inventory (BLRI) ratings and utterance-level empathy metrics, validating psychological fidelity. We also build a public Sentient Leaderboard covering 18 commercial and open-source models that uncovers substantial gaps (up to 4x) between frontier systems (GPT-4o-Latest, Gemini2.5-Pro) and earlier baselines, gaps not reflected in conventional leaderboards (e.g., Arena). SAGE thus provides a principled, scalable and interpretable tool for tracking progress toward genuinely empathetic and socially adept language agents.

cs.CL↗

arXiv2025

Disentangling Length Bias In Preference Learning Via Response-Conditioned Modeling

Reinforcement Learning from Human Feedback (RLHF) has achieved considerable success in aligning large language models (LLMs) by modeling human preferences with a learnable reward model and employing a reinforcement learning algorithm to maximize the reward model's scores. However, these reward models are susceptible to exploitation through various superficial confounding factors, with length bias emerging as a particularly significant concern. Moreover, while the pronounced impact of length bias on preference modeling suggests that LLMs possess an inherent sensitivity to length perception, our preliminary investigations reveal that fine-tuned LLMs consistently struggle to adhere to explicit length instructions. To address these two limitations, we propose a novel framework wherein the reward model explicitly differentiates between human semantic preferences and response length requirements. Specifically, we introduce a $\textbf{R}$esponse-$\textbf{c}$onditioned $\textbf{B}$radley-$\textbf{T}$erry (Rc-BT) model that enhances the model's capability in length bias mitigating and length instruction following, through training on our augmented dataset. Furthermore, we propose the Rc-RM and Rc-DPO algorithm to leverage the Rc-BT model for reward modeling and direct policy optimization (DPO) of LLMs, simultaneously mitigating length bias and promoting adherence to length instructions. Extensive experiments across various foundational models and datasets demonstrate the effectiveness and generalizability of our approach.

cs.LG↗

arXiv2025

DGRO: Enhancing LLM Reasoning via Exploration-Exploitation Control and Reward Variance Management

Inference scaling further accelerates Large Language Models (LLMs) toward Artificial General Intelligence (AGI), with large-scale Reinforcement Learning (RL) to unleash long Chain-of-Thought reasoning. Most contemporary reasoning approaches usually rely on handcrafted rule-based reward functions. However, the tarde-offs of exploration and exploitation in RL algorithms involves multiple complex considerations, and the theoretical and empirical impacts of manually designed reward functions remain insufficiently explored. In this paper, we propose Decoupled Group Reward Optimization (DGRO), a general RL algorithm for LLM reasoning. On the one hand, DGRO decouples the traditional regularization coefficient into two independent hyperparameters: one scales the policy gradient term, and the other regulates the distance from the sampling policy. This decoupling not only enables precise control over balancing exploration and exploitation, but also can be seamlessly extended to Online Policy Mirror Descent (OPMD) algorithms in Kimi k1.5 and Direct Reward Optimization. On the other hand, we observe that reward variance significantly affects both convergence speed and final model performance. We conduct both theoretical analysis and extensive empirical validation to assess DGRO, including a detailed ablation study that investigates its performance and optimization dynamics. Experimental results show that DGRO achieves state-of-the-art performance on the Logic dataset with an average accuracy of 96.9\%, and demonstrates strong generalization across mathematical benchmarks.

cs.LG↗

arXiv2025

SCoRE: Benchmarking Long-Chain Reasoning in Commonsense Scenarios

Currently, long-chain reasoning remains a key challenge for large language models (LLMs) because natural texts lack sufficient explicit reasoning data. However, existing benchmarks suffer from limitations such as narrow coverage, short reasoning paths, or high construction costs. We introduce SCoRE (Scenario-based Commonsense Reasoning Evaluation), a benchmark that synthesizes multi-hop questions from scenario schemas of entities, relations, and logical rules to assess long-chain commonsense reasoning. SCoRE contains 100k bilingual (Chinese-English) multiple-choice questions whose reasoning chains span 2-11 hops and are grouped into various difficulty levels. Each question is accompanied by fine-grained knowledge labels, explicit reasoning chains, and difficulty levels for diagnostic evaluation. Evaluation results on cutting-edge LLMs such as o3-mini and Deepseek R1 shows that even the best model attains only 69.78% accuracy on SCoRE (even only 47.91% on the hard set), with errors often stemming from rare knowledge, logical inconsistency, and over-interpretation of simple questions. SCoRE offers a scalable, extensible framework for evaluating and diagnosing the long-chain commonsense reasoning abilities of LLMs and guiding future advances in model design and training.

cs.CL↗

arXiv2025

TRENDY: Gene Regulatory Network Inference Enhanced by Transformer

Gene regulatory networks (GRNs) play a crucial role in the control of cellular functions. Numerous methods have been developed to infer GRNs from gene expression data, including mechanism-based approaches, information-based approaches, and more recent deep learning techniques, the last of which often overlook the underlying gene expression mechanisms. In this work, we introduce TRENDY, a novel GRN inference method that integrates transformer models to enhance the mechanism-based WENDY approach. Through testing on both simulated and experimental datasets, TRENDY demonstrates superior performance compared to existing methods. Furthermore, we apply this transformer-based approach to three additional inference methods, showcasing its broad potential to enhance GRN inference.

q-bio.MN↗

arXiv2025

Model-Based Closed-Loop Control Algorithm for Stochastic Partial Differential Equation Control

Neural operators have demonstrated promise in modeling and controlling systems governed by Partial Differential Equations (PDEs). Beyond PDEs, Stochastic Partial Differential Equations (SPDEs) play a critical role in modeling systems influenced by randomness, with applications in finance, physics, and beyond. However, controlling SPDE-governed systems remains a significant challenge. On the one hand, the regularity of the system's state (which can be intuitively understood as smoothness) deteriorates, making modeling and generalization more challenging. On the other hand, this stochasticity also renders control more unstable and thus less accurate. To address this gap, we propose the Model-Based Closed-Loop Control Algorithm (MB-CC), the first model-based closed-loop control method for SPDEs. MB-CC introduces two key innovations to enhance control robustness and efficiency: a Regularity Feature (RF) block and a closed-loop strategy with an operator-encoded policy network. The RF block, inspired by the regularity structure theory of SPDEs, addresses noise-induced irregularities by transforming the network's input, including the system state and noise-perturbed external forces, into a refined feature space for improved forward prediction. Compared to previous works using regularity features, we introduce a new parameterization, data augmentation, and extend the RF block as a plug-and-play component. Additionally, to achieve closed-loop control, we introduce an operator-encoded policy network to map the current state to optimal control, which integrates physical priors and swiftly makes decisions based on states returned by the environment. We conduct a systematic evaluation of MB-CC on two notable SPDEs, showcasing its effectiveness and efficiency. The ablation studies show its ability to handle stochasticity more effectively.

eess.SY↗

arXiv2025

MSCI: Addressing CLIP's Inherent Limitations for Compositional Zero-Shot Learning

Compositional Zero-Shot Learning (CZSL) aims to recognize unseen state-object combinations by leveraging known combinations. Existing studies basically rely on the cross-modal alignment capabilities of CLIP but tend to overlook its limitations in capturing fine-grained local features, which arise from its architectural and training paradigm. To address this issue, we propose a Multi-Stage Cross-modal Interaction (MSCI) model that effectively explores and utilizes intermediate-layer information from CLIP's visual encoder. Specifically, we design two self-adaptive aggregators to extract local information from low-level visual features and integrate global information from high-level visual features, respectively. These key information are progressively incorporated into textual representations through a stage-by-stage interaction mechanism, significantly enhancing the model's perception capability for fine-grained local visual information. Additionally, MSCI dynamically adjusts the attention weights between global and local visual information based on different combinations, as well as different elements within the same combination, allowing it to flexibly adapt to diverse scenarios. Experiments on three widely used datasets fully validate the effectiveness and superiority of the proposed model. Data and code are available at https://github.com/ltpwy/MSCI.

cs.CV↗

arXiv2025

Pitch Angle Measurement Method based on Detector Counts Distribution. -I. Basic conception

As an X-ray and gamma-ray all-sky monitor aiming for high energy astrophysical transients, Gravitational-wave high-energy Electromagnetic Counterpart All-sky Monitor (GECAM) has also made a series of observational discoveries on burst events of gamma-rays and particles in the low Earth orbit. Pitch angle is one of the key parameters of charged particles traveling around geomagnetic field. However, the usage of the GECAM-style instruments to measure the pitch angle of charged particles is still lacking. Here we propose a novel method for GECAM and similar instruments to measure the pitch angle of charged particles based on detector counts distribution. The basic conception of this method and simulation studies are described. With this method, the pitch angle of a peculiar electron precipitation event detected by GECAM-C is derived to be about 90$^\circ$, demonstrating the feasibility of our method. We note that the application of this method on GECAM-style instruments may open a new window for studying space particle events, such as Terrestrial Electron Beams (TEBs) and Lightning-induced Electron Precipitations (LEPs).

astro-ph.IM↗

arXiv2025

GECAM Discovery of Peculiar Oscillating Particle Precipitation Events

Charged particle precipitation typically manifests as a gradual increase and decrease of flux observed by space detectors. Cases with rapidly flux variation are very rare. Periodic events are even more extraordinary. These oscillating particle precipitation (OPP) events are usually attributed to the bounce motion of electrons, which are induced by lightning. Owing to the observation limitations, there has been debate regarding whether these oscillations originate from temporal flux evolution or spatial structure evolution. Here we report three peculiar charged particle precipitation events detected by GECAM during a geomagnetic storm on March 21, 2024, with two exhibiting significant periodicity. These events were observed around the same region during three consecutive orbits. Through comprehensive temporal and spectral analyses, we revealed that one of the OPP events exhibited a transition in spectral lag of mini-pulses, shifting from "softer-earlier" to "softer-later" while showing no significant time evolution in overall frequency characteristics. And there is no association found between these two OPP events and lightning activity. Several possible scenarios are discussed to explain these charged particles with a life time of more than 3.5 hours, but the nature of these three events remains an enigma. We suggest that these GECAM-detected OPP events may represent a new type of particle precipitation event or a peculiar Lightning-induced Electron Precipitations (LEPs).

astro-ph.HE↗

arXiv2025

Deformable Beta Splatting

3D Gaussian Splatting (3DGS) has advanced radiance field reconstruction by enabling real-time rendering. However, its reliance on Gaussian kernels for geometry and low-order Spherical Harmonics (SH) for color encoding limits its ability to capture complex geometries and diverse colors. We introduce Deformable Beta Splatting (DBS), a deformable and compact approach that enhances both geometry and color representation. DBS replaces Gaussian kernels with deformable Beta Kernels, which offer bounded support and adaptive frequency control to capture fine geometric details with higher fidelity while achieving better memory efficiency. In addition, we extended the Beta Kernel to color encoding, which facilitates improved representation of diffuse and specular components, yielding superior results compared to SH-based methods. Furthermore, Unlike prior densification techniques that depend on Gaussian properties, we mathematically prove that adjusting regularized opacity alone ensures distribution-preserved Markov chain Monte Carlo (MCMC), independent of the splatting kernel type. Experimental results demonstrate that DBS achieves state-of-the-art visual quality while utilizing only 45% of the parameters and rendering 1.5x faster than 3DGS-MCMC, highlighting the superior performance of DBS for real-time radiance field rendering. Interactive demonstrations and source code are available on our project website: https://rongliu-leo.github.io/beta-splatting/.

cs.CV↗

arXiv2025

Matrix Healy Plot: A Practical Tool for Visual Assessment of Matrix-Variate Normality

Matrix-valued data, where each observation is represented as a matrix, frequently arises in various scientific disciplines. Modeling such data often relies on matrix-variate normal distributions, making matrix-variate normality testing crucial for valid statistical inference. Recently, the Distance-Distance (DD) plot has been introduced as a graphical tool for visually assessing matrix-variate normality. However, the Mahalanobis squared distances (MSD) used in the DD plot require vectorizing matrix observations, restricting its applicability to cases where the dimension of the vectorized data does not exceed the sample size. To address this limitation, we propose a novel graphical method called the Matrix Healy (MHealy) plot, an extension of the Healy plot for vector-valued data. This new plot is based on more accurate matrix-based MSD that leverages the inherent structure of matrix data. Consequently, it offers a more reliable visual assessment. Importantly, the MHealy plot eliminates the sample size restriction of the DD plot and hence more applicable to matrix-valued data. Empirical results demonstrate its effectiveness and practicality compared to the DD plot across various scenarios, particularly in cases where the DD plot is not available due to limited sample sizes.

stat.ME↗

arXiv2025

Stable self-charged perovskite quantum rods for liquid laser with near-zero threshold

Colloidal quantum dots (QDs) are promising optical gain materials that require further threshold reduction to realize their full potential. While QD charging theoretically reduces the threshold to zero, its effectiveness has been limited by strong Auger recombination and unstable charging. Here we theoretically reveal the optimal combination of charging number and Auger recombination to minimize the lasing threshold. Experimentally, we develop stable self-charged perovskite quantum rods (QRs) as an alternative to QDs via state engineering and Mn-doping strategy. An unprecedented two-order-of-magnitude reduction in nonradiative Auger recombination enables QRs to support a sufficient charging number of up to 6. The QR liquid lasing is then achieved with a near-zero threshold of 0.098 using quasi-continuous pumping of nanosecond pulses, which is the lowest threshold among all reported QD lasers. These achievements demonstrate the potential of the specially engineered QRs as an excellent gain media and pave the way for their prospective applications.

cond-mat.mes-hall↗

arXiv2025

HMCF: A Human-in-the-loop Multi-Robot Collaboration Framework Based on Large Language Models

Rapid advancements in artificial intelligence (AI) have enabled robots to performcomplex tasks autonomously with increasing precision. However, multi-robot systems (MRSs) face challenges in generalization, heterogeneity, and safety, especially when scaling to large-scale deployments like disaster response. Traditional approaches often lack generalization, requiring extensive engineering for new tasks and scenarios, and struggle with managing diverse robots. To overcome these limitations, we propose a Human-in-the-loop Multi-Robot Collaboration Framework (HMCF) powered by large language models (LLMs). LLMs enhance adaptability by reasoning over diverse tasks and robot capabilities, while human oversight ensures safety and reliability, intervening only when necessary. Our framework seamlessly integrates human oversight, LLM agents, and heterogeneous robots to optimize task allocation and execution. Each robot is equipped with an LLM agent capable of understanding its capabilities, converting tasks into executable instructions, and reducing hallucinations through task verification and human supervision. Simulation results show that our framework outperforms state-of-the-art task planning methods, achieving higher task success rates with an improvement of 4.76%. Real-world tests demonstrate its robust zero-shot generalization feature and ability to handle diverse tasks and environments with minimal human intervention.

cs.RO↗

arXiv2025

Residual-based Alternative Partial Least Squares for Generalized Functional Linear Models

Many biomedical studies collect high-dimensional medical imaging data to identify biomarkers for the detection, diagnosis, and treatment of human diseases. Consequently, it is crucial to develop accurate models that can predict a wide range of clinical outcomes (both discrete and continuous) based on imaging data. By treating imaging predictors as functional data, we propose a residual-based alternative partial least squares (RAPLS) model for a broad class of generalized functional linear models that incorporate both functional and scalar covariates. Our RAPLS method extends the alternative partial least squares (APLS) algorithm iteratively to accommodate additional scalar covariates and non-continuous outcomes. We establish the convergence rate of the RAPLS estimator for the unknown slope function and, with an additional calibration step, we prove the asymptotic normality and efficiency of the calibrated RAPLS estimator for the scalar parameters. The effectiveness of the RAPLS algorithm is demonstrated through multiple simulation studies and an application predicting Alzheimer's disease progression using neuroimaging data from the Alzheimer's Disease Neuroimaging Initiative (ADNI).

stat.ME↗

arXiv2025

GenCLS++: Pushing the Boundaries of Generative Classification in LLMs Through Comprehensive SFT and RL Studies Across Diverse Datasets

As a fundamental task in machine learning, text classification plays a crucial role in many areas. With the rapid scaling of Large Language Models (LLMs), particularly through reinforcement learning (RL), there is a growing need for more capable discriminators. Consequently, advances in classification are becoming increasingly vital for enhancing the overall capabilities of LLMs. Traditional discriminative methods map text to labels but overlook LLMs' intrinsic generative strengths. Generative classification addresses this by prompting the model to directly output labels. However, existing studies still rely on simple SFT alone, seldom probing the interplay between training and inference prompts, and no work has systematically leveraged RL for generative text classifiers and unified SFT, RL, and inference-time prompting in one framework. We bridge this gap with GenCLS++, a framework that jointly optimizes SFT and RL while systematically exploring five high-level strategy dimensions-in-context learning variants, category definitions, explicit uncertainty labels, semantically irrelevant numeric labels, and perplexity-based decoding-during both training and inference. After an SFT "policy warm-up," we apply RL with a simple rule-based reward, yielding sizable extra gains. Across seven datasets, GenCLS++ achieves an average accuracy improvement of 3.46% relative to the naive SFT baseline; on public datasets, this improvement rises to 4.00%. Notably, unlike reasoning-intensive tasks that benefit from explicit thinking processes, we find that classification tasks perform better without such reasoning steps. These insights into the role of explicit reasoning provide valuable guidance for future LLM applications.

cs.CL↗

arXiv2025

RoboVerse: Towards a Unified Platform, Dataset and Benchmark for Scalable and Generalizable Robot Learning

Data scaling and standardized evaluation benchmarks have driven significant advances in natural language processing and computer vision. However, robotics faces unique challenges in scaling data and establishing evaluation protocols. Collecting real-world data is resource-intensive and inefficient, while benchmarking in real-world scenarios remains highly complex. Synthetic data and simulation offer promising alternatives, yet existing efforts often fall short in data quality, diversity, and benchmark standardization. To address these challenges, we introduce RoboVerse, a comprehensive framework comprising a simulation platform, a synthetic dataset, and unified benchmarks. Our simulation platform supports multiple simulators and robotic embodiments, enabling seamless transitions between different environments. The synthetic dataset, featuring high-fidelity physics and photorealistic rendering, is constructed through multiple approaches. Additionally, we propose unified benchmarks for imitation learning and reinforcement learning, enabling evaluation across different levels of generalization. At the core of the simulation platform is MetaSim, an infrastructure that abstracts diverse simulation environments into a universal interface. It restructures existing simulation environments into a simulator-agnostic configuration system, as well as an API aligning different simulator functionalities, such as launching simulation environments, loading assets with initial states, stepping the physics engine, etc. This abstraction ensures interoperability and extensibility. Comprehensive experiments demonstrate that RoboVerse enhances the performance of imitation learning, reinforcement learning, world model learning, and sim-to-real transfer. These results validate the reliability of our dataset and benchmarks, establishing RoboVerse as a robust solution for advancing robot learning.

cs.RO↗

arXiv2025

Insight-HXMT observations of the extremely bright GRB 221009A

The Hard X-ray Modulation Telescope (\insight) detected GRB 221009A, the brightest gamma-ray burst observed to date, with all its three telescopes, i.e. High Energy telescope (HE, 20-250 keV), Medium Energy telescope (ME, 5-30 keV), and Low Energy telescope (LE, 1-10 keV). Here we present the detailed observation results of all three telescopes of \insight~ on the prompt emission of GRB 221009A. After dead-time and data saturation correction, we recovered the light curves of HE, ME and LE telescopes and find that they generally track the GECAM-C low gain light curves that are free of data saturation issues. Particularly, the ME light curve matches the GECAM-C light curve in low gain mode above 400 keV, while the LE light curve is more consistent with the GECAM-C above 1.5 MeV. Based on simulation, we find that the signals recorded by the ME and LE are actually caused by the secondary particles produced by the interaction between GRB gamma-ray photons and the material of the satellite. Interestingly, the consistency between ME and LE light curves and GECAM-C demonstrates that ME and LE data could be used to characterize the GRB properties. Espeically, the high time resolution light curve of ME allowed us, for the first time, to calculate the minimum variability timescale (MVT = 0.10 s) of the main burst episode of GRB 221009A.

astro-ph.HE↗

arXiv2025

Symbolic Representation for Any-to-Any Generative Tasks

We propose a symbolic generative task description language and a corresponding inference engine capable of representing arbitrary multimodal tasks as structured symbolic flows. Unlike conventional generative models that rely on large-scale training and implicit neural representations to learn cross-modal mappings, often at high computational cost and with limited flexibility, our framework introduces an explicit symbolic representation comprising three core primitives: functions, parameters, and topological logic. Leveraging a pre-trained language model, our inference engine maps natural language instructions directly to symbolic workflows in a training-free manner. Our framework successfully performs over 12 diverse multimodal generative tasks, demonstrating strong performance and flexibility without the need for task-specific tuning. Experiments show that our method not only matches or outperforms existing state-of-the-art unified models in content quality, but also offers greater efficiency, editability, and interruptibility. We believe that symbolic task representations provide a cost-effective and extensible foundation for advancing the capabilities of generative AI.

cs.LG↗