arXiv subjects

Guo Yu

Publications and source records attributed to Guo Yu.

At least 19 recordsLinked to original sources

arXiv2026

RSRA: Training-Free Probing of Representation Sensitivity for Efficient LoRA Rank Allocation

Parameter-efficient fine-tuning enables large language models to adapt to downstream tasks with substantially lower computational and storage cost, and Low-Rank Adaptation (LoRA) is among its most widely used techniques. However, vanilla LoRA assigns a uniform rank to all adapted modules, while existing adaptive methods either incur additional optimization overhead or rely on static weights and local gradients that do not capture task-conditioned representation changes. We propose RSRA, a training-free rank allocator that estimates where adaptation capacity is most needed through forward-only representation sensitivity probing on a small calibration set. Specifically, RSRA uses Spectral Effective Rank to allocate capacity across layers, measures module-wise hidden-state displacement under standardized virtual low-rank updates with the Frechet Distance, and combines both signals through hierarchical normalization to produce a task-aware rank configuration before fine-tuning. Across commonsense reasoning and natural language understanding benchmarks with Qwen3-4B and Mistral-7B, RSRA achieves the highest average performance in all three reported model-benchmark settings and a 1.48x-1.93x speedup in allocation time over the fastest competing pre-allocation method. When integrated with DoRA, LoRA-FA, and PiSSA, RSRA improves 15 of the 18 evaluated combinations and increases the average performance of all three PEFT methods.

cs.CL↗

arXiv2026

Moment-Based Selection of Multiresponse Linear Mixed-Effects Models

We propose MOMENT (\textbf{MO}ment-Based \textbf{M}ixed-\textbf{E}ffects Selectio\textbf{N} and Es\textbf{T}imation), a stage-wise moment-based framework that exploits second-order cross-moment identities to select and estimate the random-effects covariance matrix and fixed-effects coefficients. By inducing sparsity through its diagonal under a positive semidefinite constraint, the random-effects selection problem reduces to a smooth constrained convex optimization problem that can be solved efficiently by projected gradient descent. We further establish finite-sample theoretical guarantees for the proposed procedure, including random-effects selection consistency and fixed-effects selection consistency under joint sub-Weibull errors. Simulation studies show that MOMENT performs competitively overall and can substantially outperform separate univariate analyses when responses are correlated. An application to the hemodialysis dataset demonstrates that the proposed method yields an interpretable and flexible approach for multivariate longitudinal data.

stat.ME↗

arXiv2026

Y-BotFrame: An Extensible Embodied Agent Framework for Quadruped Robot Assistants

Quadruped robots are capable of traversing a wide range of complex terrains with high flexibility. As highly mobile ground-based intelligent platforms, they can be equipped with modules for navigation control, environmental perception, and intelligent interaction, thereby serving as real-world mobile deployment platforms for various algorithms. In this paper, we introduce Y-BotFrame, an extensible embodied platform that turns a robot into an intelligent ground assistant. Y-BotFrame integrates multimodal perception capabilities, including speech, vision, and LiDAR, and employs a large language model as the cognitive core for environmental understanding, contextual reasoning, and task planning. The system maps user natural-language instructions into executable embodied task units that can be carried out by the robot. Y-BotFrame supports natural interaction through voice commands and visual feedback, removing the need for a remote controller and enabling efficient human-robot collaboration. With a highly extensible framework, Y-BotFrame supports plug-and-play integration of new functional modules as well as modular upgrades and iterative development, offering a reference implementation for the real-world deployment of general-purpose, instruction-driven embodied agents.The supplementary video is available at https://xdei-group.github.io/Y-BotFrame/.

cs.RO↗

arXiv2026

Dense Supervision, Sparse Updates: On the Sparsity and Geometry of On-Policy Distillation

On-policy distillation (OPD) has recently become a prominent post-training recipe by combining two desirable ingredients: on-policy student-generated trajectories and dense token-level teacher supervision. Yet how this hybrid training regime shapes a model remains poorly understood. We characterize the sparsity and geometry of OPD parameter updates across several language and vision-language model pairs and application settings. OPD updates are small and coordinate-sparse at checkpoint precision, while remaining distributed across layers and modules. This sparse support is operationally meaningful: masked training on the discovered subnetwork nearly recovers full-training performance. At the matrix level, the updates are numerically full-rank but spectrally concentrated. Their visible supports avoid coordinates emphasized by the source's principal structure and favor low-magnitude source coordinates, while the source singular-value spectra change little. Together, these findings show that OPD exhibits important weight-space signatures of on-policy post-training despite using dense teacher supervision.

cs.LG↗

arXiv2026

AerialClaw: An Open-Source Framework for LLM-Driven Autonomous Aerial Agents

Unmanned aerial vehicles (UAVs) are increasingly used in inspection, search and rescue, environmental monitoring, and emergency response. However, most UAV applications still rely on pre-defined command sequences or task-specific pipelines, where developers manually connect perception, planning, flight control, simulation, logging, and safety modules. This limits the flexibility, reproducibility, and extensibility of autonomous aerial systems. This paper presents AerialClaw, an open-source software framework that enables UAVs to operate as decision-making aerial agents rather than merely command-following platforms. Given a natural-language mission, AerialClaw allows an LLM-based agent to understand the task, maintain context, invoke executable aerial skills, observe perception and runtime feedback, and iteratively update its decisions in a closed loop. The framework adopts a modular brain-skill-runtime architecture, combining hard skills for atomic UAV operations, Markdown-based soft skills for reusable task strategies, document-driven agent state and capability boundaries, memory-driven reflection, safety-oriented runtime validation, and platform-agnostic execution adapters. AerialClaw supports lightweight mock execution, PX4 SITL with Gazebo, and AirSim-based simulation, together with a web console, pluggable model backends, example missions, simulation assets, and staged deployment scripts. By combining standardized aerial skills, document-driven agent state, memory, and closed-loop LLM decision-making, AerialClaw provides a reproducible and extensible open-source framework for building UAV systems that can interpret missions, make decisions, execute skills, and adapt their behavior from feedback.

cs.RO↗

arXiv2026

Benchmarking Egocentric Clinical Intent Understanding Capability for Medical Multimodal Large Language Models

Medical Multimodal Large Language Models (Med-MLLMs) require egocentric clinical intent understanding for real-world deployment, yet existing benchmarks fail to evaluate this critical capability. To address these challenges, we introduce MedGaze-Bench, the first benchmark leveraging clinician gaze as a Cognitive Cursor to assess intent understanding across surgery, emergency simulation, and diagnostic interpretation. Our benchmark addresses three fundamental challenges: visual homogeneity of anatomical structures, strict temporal-causal dependencies in clinical workflows, and implicit adherence to safety protocols. We propose a Three-Dimensional Clinical Intent Framework evaluating: (1) Spatial Intent: discriminating precise targets amid visual noise, (2) Temporal Intent: inferring causal rationale through retrospective and prospective reasoning, and (3) Standard Intent: verifying protocol compliance through safety checks. Beyond accuracy metrics, we introduce Trap QA mechanisms to stress-test clinical reliability by penalizing hallucinations and cognitive sycophancy. Experiments reveal current MLLMs struggle with egocentric intent due to over-reliance on global features, leading to fabricated observations and uncritical acceptance of invalid instructions.

cs.CV↗

arXiv2025

Mamba4Net: Distilled Hybrid Mamba Large Language Models For Networking

Transformer-based large language models (LLMs) are increasingly being adopted in networking research to address domain-specific challenges. However, their quadratic time complexity and substantial model sizes often result in significant computational overhead and memory constraints, particularly in resource-constrained environments. Drawing inspiration from the efficiency and performance of the Deepseek-R1 model within the knowledge distillation paradigm, this paper introduces Mamba4Net, a novel cross-architecture distillation framework. Mamba4Net transfers networking-specific knowledge from transformer-based LLMs to student models built on the Mamba architecture, which features linear time complexity. This design substantially enhances computational efficiency compared to the quadratic complexity of transformer-based models, while the reduced model size further minimizes computational demands, improving overall performance and resource utilization. To evaluate its effectiveness, Mamba4Net was tested across three diverse networking tasks: viewport prediction, adaptive bitrate streaming, and cluster job scheduling. Compared to existing methods that do not leverage LLMs, Mamba4Net demonstrates superior task performance. Furthermore, relative to direct applications of transformer-based LLMs, it achieves significant efficiency gains, including a throughput 3.96 times higher and a storage footprint of only 5.48% of that required by previous LLM-based approaches. These results highlight Mamba4Net's potential to enable the cost-effective application of LLM-derived knowledge in networking contexts. The source code is openly available to support further research and development.

cs.NI↗

arXiv2025

Pressure in active matter

In the last decade, the study of pressure in active matter has attracted growing attention due to its fundamental relevance to nonequilibrium statistical physics. Active matter systems are composed of particles that consume energy to sustain persistent motion, which are inherently far from equilibrium. These particles can exhibit complex behaviors, including motility-induced phase separation, density-dependent clustering, and anomalous stress distributions, motivating the introduction of active swim stress and swim pressure. Unlike in passive fluids, pressure in active systems emerges from momentum flux originated from swim force rather than equilibrium conservative interactions, offering a distinct perspective for understanding their mechanical response. Simple models of active Brownian particles (ABPs) have been employed in theoretical and simulation studies across both dilute and dense regimes, revealing that pressure is a state function and exhibits a nontrivial dependence on density. Together with nonequilibrium statistical concepts such as effective temperature and effective adhesion, pressure offers important insight for understanding behaviors in active matter such as sedimentation equilibrium and motility induced phase separation. Extensions of ABPs models beyond their simplest form have underscored the fragility of pressure-based equation of state, which can break down under factors such as density-dependent velocity, torque, complex boundary geometries and interactions. Building on these developments, this review provides a comprehensive survey of theoretical and experimental advances, with particular emphasis on the microscopic origins of active pressure and the mechanisms underlying the breakdown of the equation of state.

cond-mat.soft↗

arXiv2025

Beyond the Leaderboard: Rethinking Medical Benchmarks for Large Language Models

Large language models (LLMs) show significant potential in healthcare, prompting numerous benchmarks to evaluate their capabilities. However, concerns persist regarding the reliability of these benchmarks, which often lack clinical fidelity, robust data management, and safety-oriented evaluation metrics. To address these shortcomings, we introduce MedCheck, the first lifecycle-oriented assessment framework specifically designed for medical benchmarks. Our framework deconstructs a benchmark's development into five continuous stages, from design to governance, and provides a comprehensive checklist of 46 medically-tailored criteria. Using MedCheck, we conducted an in-depth empirical evaluation of 53 medical LLM benchmarks. Our analysis uncovers widespread, systemic issues, including a profound disconnect from clinical practice, a crisis of data integrity due to unmitigated contamination risks, and a systematic neglect of safety-critical evaluation dimensions like model robustness and uncertainty awareness. Based on these findings, MedCheck serves as both a diagnostic tool for existing benchmarks and an actionable guideline to foster a more standardized, reliable, and transparent approach to evaluating AI in healthcare.

cs.CL↗

arXiv2025

Where and How to Enhance: Discovering Bit-Width Contribution for Mixed Precision Quantization

Mixed precision quantization (MPQ) is an effective quantization approach to achieve accuracy-complexity trade-off of neural network, through assigning different bit-widths to network activations and weights in each layer. The typical way of existing MPQ methods is to optimize quantization policies (i.e., bit-width allocation) in a gradient descent manner, termed as Differentiable (DMPQ). At the end of the search, the bit-width associated to the quantization parameters which has the largest value will be selected to form the final mixed precision quantization policy, with the implicit assumption that the values of quantization parameters reflect the operation contribution to the accuracy improvement. While much has been discussed about the MPQ improvement, the bit-width selection process has received little attention. We study this problem and argue that the magnitude of quantization parameters does not necessarily reflect the actual contribution of the bit-width to the task performance. Then, we propose a Shapley-based MPQ (SMPQ) method, which measures the bit-width operation direct contribution on the MPQ task. To reduce computation cost, a Monte Carlo sampling-based approximation strategy is proposed for Shapley computation. Extensive experiments on mainstream benchmarks demonstrate that our SMPQ consistently achieves state-of-the-art performance than gradient-based competitors.

cs.LG↗

arXiv2025

Reluctant Interaction Inference after Additive Modeling

Additive models enjoy the flexibility of nonlinear models while still being readily understandable to humans. By contrast, other nonlinear models, which involve interactions between features, are not only harder to fit but also substantially more complicated to explain. Guided by the principle of parsimony, a data analyst therefore may naturally be reluctant to move beyond an additive model unless it is truly warranted. To put this principle of interaction reluctance into practice, we formulate the problem as a hypothesis test with a fitted sparse additive model (SPAM) serving as the null. Because our hypotheses on interaction effects are formed after fitting a SPAM to the data, we adopt a selective inference approach to construct p-values that properly account for this data adaptivity. Our approach makes use of external randomization to obtain the distribution of test statistics conditional on the SPAM fit, allowing us to derive valid p-values, corrected for the over-optimism introduced by the data-adaptive process prior to the test. Through experiments on simulated and real data, we illustrate that--even with small amounts of external randomization--this rigorous modeling approach enjoys considerable advantages over naive methods and data splitting.

stat.ME↗

arXiv2025

PolyBERT: Fine-Tuned Poly Encoder BERT-Based Model for Word Sense Disambiguation

Mainstream Word Sense Disambiguation (WSD) approaches have employed BERT to extract semantics from both context and definitions of senses to determine the most suitable sense of a target word, achieving notable performance. However, there are two limitations in these approaches. First, previous studies failed to balance the representation of token-level (local) and sequence-level (global) semantics during feature extraction, leading to insufficient semantic representation and a performance bottleneck. Second, these approaches incorporated all possible senses of each target word during the training phase, leading to unnecessary computational costs. To overcome these limitations, this paper introduces a poly-encoder BERT-based model with batch contrastive learning for WSD, named PolyBERT. Compared with previous WSD methods, PolyBERT has two improvements: (1) A poly-encoder with a multi-head attention mechanism is utilized to fuse token-level (local) and sequence-level (global) semantics, rather than focusing on just one. This approach enriches semantic representation by balancing local and global semantics. (2) To avoid redundant training inputs, Batch Contrastive Learning (BCL) is introduced. BCL utilizes the correct senses of other target words in the same batch as negative samples for the current target word, which reduces training inputs and computational cost. The experimental results demonstrate that PolyBERT outperforms baseline WSD methods such as Huang's GlossBERT and Blevins's BEM by 2\% in F1-score. In addition, PolyBERT with BCL reduces GPU hours by 37.6\% compared with PolyBERT without BCL.

cs.AI↗

arXiv2025

UrbanCraft: Urban View Extrapolation via Hierarchical Sem-Geometric Priors

Existing neural rendering-based urban scene reconstruction methods mainly focus on the Interpolated View Synthesis (IVS) setting that synthesizes from views close to training camera trajectory. However, IVS can not guarantee the on-par performance of the novel view outside the training camera distribution (\textit{e.g.}, looking left, right, or downwards), which limits the generalizability of the urban reconstruction application. Previous methods have optimized it via image diffusion, but they fail to handle text-ambiguous or large unseen view angles due to coarse-grained control of text-only diffusion. In this paper, we design UrbanCraft, which surmounts the Extrapolated View Synthesis (EVS) problem using hierarchical sem-geometric representations serving as additional priors. Specifically, we leverage the partially observable scene to reconstruct coarse semantic and geometric primitives, establishing a coarse scene-level prior through an occupancy grid as the base representation. Additionally, we incorporate fine instance-level priors from 3D bounding boxes to enhance object-level details and spatial relationships. Building on this, we propose the \textbf{H}ierarchical \textbf{S}emantic-Geometric-\textbf{G}uided Variational Score Distillation (HSG-VSD), which integrates semantic and geometric constraints from pretrained UrbanCraft2D into the score distillation sampling process, forcing the distribution to be consistent with the observable scene. Qualitative and quantitative comparisons demonstrate the effectiveness of our methods on EVS problem.

cs.CV↗

arXiv2025

Unconventional Superconducting Phase Diagram of Monolayer WTe2

The existence of a quantum critical point (QCP) and fluctuations around it are believed to be important for understanding the phase diagram in unconventional superconductors such as cuprates, iron pnictides, and heavy fermion superconductors. However, the QCP is usually buried deep within the superconducting dome and is difficult to investigate. The connection between quantum critical fluctuations and superconductivity remains an outstanding problem in condensed matter. Here combining both electrical transport and Nernst experiments, we explicitly demonstrate the onset of superconductivity at an unconventional QCP in gate-tuned monolayer tungsten ditelluride (WTe2), with features incompatible with the conventional Bardeen-Cooper-Schrieffer (BCS) scenario. The results lead to a novel superconducting phase diagram that is distinguished from other known superconductors. Two distinct gate-tuned quantum phase transitions are observed at the ends of the superconducting dome. We find that quantum fluctuations around the QCP of the underdoped regime are essential for understanding how the monolayer superconductivity is established. The unconventional phase diagram we report here illustrates a previously unknown relation between superconductivity and QCP.

cond-mat.mes-hall↗

arXiv2024

GSGTrack: Gaussian Splatting-Guided Object Pose Tracking from RGB Videos

Tracking the 6DoF pose of unknown objects in monocular RGB video sequences is crucial for robotic manipulation. However, existing approaches typically rely on accurate depth information, which is non-trivial to obtain in real-world scenarios. Although depth estimation algorithms can be employed, geometric inaccuracy can lead to failures in RGBD-based pose tracking methods. To address this challenge, we introduce GSGTrack, a novel RGB-based pose tracking framework that jointly optimizes geometry and pose. Specifically, we adopt 3D Gaussian Splatting to create an optimizable 3D representation, which is learned simultaneously with a graph-based geometry optimization to capture the object's appearance features and refine its geometry. However, the joint optimization process is susceptible to perturbations from noisy pose and geometry data. Thus, we propose an object silhouette loss to address the issue of pixel-wise loss being overly sensitive to pose noise during tracking. To mitigate the geometric ambiguities caused by inaccurate depth information, we propose a geometry-consistent image pair selection strategy, which filters out low-confidence pairs and ensures robust geometric optimization. Extensive experiments on the OnePose and HO3D datasets demonstrate the effectiveness of GSGTrack in both 6DoF pose tracking and object reconstruction.

cs.CV↗

arXiv2024

Convex Estimation of Gaussian Graphical Regression Models with Covariates

Gaussian graphical models (GGMs) are widely used to recover the conditional independence structure among random variables. Recent work has sought to incorporate auxiliary covariates to improve estimation, particularly in applications such as co-expression quantitative trait locus (eQTL) studies, where both gene expression levels and their conditional dependence structure may be influenced by genetic variants. Existing approaches to covariate-adjusted GGMs either restrict covariate effects to the mean structure or lead to nonconvex formulations when jointly estimating the mean and precision matrix. In this paper, we propose a convex framework that simultaneously estimates the covariate-adjusted mean and precision matrix via a natural parametrization of the multivariate Gaussian likelihood. The resulting formulation enables joint convex optimization and yields improved theoretical guarantees under high-dimensional scaling, where the sparsity and dimension of covariates grow with the sample size. We support our theoretical findings with numerical simulations and demonstrate the practical utility of the proposed method through a reanalysis of an eQTL study of glioblastoma multiforme and an analysis of diet on the human gut microbiome.

stat.ME↗

arXiv2024

Anomalous Superconductivity in Twisted MoTe2 Nanojunctions

Introducing superconductivity in topological materials can lead to innovative electronic phases and device functionalities. Here, we present a new strategy for quantum engineering of superconducting junctions in moire materials through direct, on-chip, and fully encapsulated 2D crystal growth. We achieve robust and designable superconductivity in Pd-metalized twisted bilayer molybdenum ditelluride (MoTe2) and observe anomalous superconducting effects in high-quality junctions across ~ 20 moire cells. Surprisingly, the junction develops enhanced, instead of weakened, superconducting behaviors, exhibiting fluctuations to a higher critical magnetic field compared to its adjacent Pd7MoTe2 superconductor. Additionally, the critical current further exhibits a striking V-shaped minimum at zero magnetic field. These features are unexpected in conventional Josephson junctions and indeed absent in junctions of natural bilayer MoTe2 created using the same approach. We discuss implications of these observations, including the possible formation of mixed even- and odd-parity superconductivity at the moire junctions. Our results also demonstrate a pathway to engineer and investigate superconductivity in fractional Chern insulators.

cond-mat.supr-con↗

arXiv2024

Large Thermopower with Sign-Alternating Quantum Oscillations in Insulating Monolayer WTe2

The detection of Landau-level-like energy structures near the chemical potential of an insulator is essential to the search for a class of correlated electronic matter hosting charge-neutral fermions and Fermi surfaces, a long-proposed concept that remains elusive experimentally. Here we introduce and demonstrate that the magneto-thermoelectric response of a quantum insulator can reveal critical information not available via other approaches. We report the observation of large thermopower together with quantum oscillations (QOs) in the hole-doped insulating state of monolayer tungsten ditelluride (WTe2) in magnetic fields. The measured low temperature magneto-thermopower exceeds k_B/e by more than an order of magnitude, where k_B is the Boltzmann constant and e the elementary charge. This large thermopower is a characteristic of an insulating state, consistent with high resistivity. However, as the magnetic field is swept, QOs develop in the thermopower, which remarkably undergoes sign-changes that mimic the quantum characteristic of metals due to Landau quantization. The resistivity of the sample is orders of magnitude larger than the value expected from the QOs. The sign-change in the thermoelectric response directly implies the presence of a field-induced Landau-level-like structure at the chemical potential of the insulator. Neither the large thermopower nor the sign-changes can be induced by the metallic gate nearby. Our results demonstrate a new dilemma for investigating low energy excitations in correlated materials featuring mixed quantum characteristics of metals and insulators.

cond-mat.str-el↗