arXiv subjects

Yida Yang

Publications and source records attributed to Yida Yang.

10 recordsLinked to original sources

arXiv2026

RoboScope Kit: Unified Interfaces for Fast and Scalable Embodied-AI Evaluation

Simulation benchmarks are the standard way to compare robot policies, but each benchmark brings its own evaluator, observation layout, and action convention, and evaluation is slow: 500 RoboTwin Easy episodes with one simulator and one model on one GPU take 359.7 minutes. Profiling the official StarVLA evaluator shows that simulation occupies CPU cores while the GPU averages 14.3% utilization. RoboScope Kit wraps each model once in a Python class called a Policy, and each benchmark once in an adapter around the benchmark's own evaluator, so the benchmark's rules and scoring stay unchanged. Scheduling runs simulator workers and Policy replicas as separate pools, sizes both from measured probes, and shards work deterministically across GPU instances; runtime optimization renders only the observations an action-chunk Policy reads. Adapters cover six benchmark families. On LIBERO-Long with a StarVLA checkpoint on one 48 GiB RTX 4090, the concurrency search selects twelve simulator workers sharing three Policy replicas, which evaluate 500 episodes 4.45$\times$ faster than the official StarVLA evaluation with equivalent success rates, 2.78$\times$ faster than RLinf, and 1.52$\times$ faster than vla-eval at its tuned configuration. With CUDA MPS, which lets the replicas run their kernels concurrently, the search selects sixteen workers sharing four replicas, 6.16$\times$ faster than StarVLA, 3.85$\times$ faster than RLinf, and 2.10$\times$ faster than vla-eval. On RoboTwin Easy, four GPUs with chunk-boundary observation complete the 500 episodes in 14.4 minutes, 25.04$\times$ faster than the single-GPU baseline, with identical success outcomes and terminal steps.

cs.RO↗

arXiv2026

Trusted Weights, Treacherous Optimizations? Optimization-Triggered Backdoor Attacks on LLMs

Inference optimization aims to minimize the latency and resource consumption of LLM inference while preserving output quality, making large-scale deployment practical and cost-effective. However, optimized execution can introduce small numerical inconsistencies from the original model. We reveal that this inconsistency not only causes the model's outputs to diverge, but more critically can introduce hidden backdoors. The backdoor remains dormant under standard unoptimized execution and is activated only when inference optimization is enabled, allowing it to evade existing backdoor detection pipelines. We first introduce the Input-Specific Optimization Backdoor (ISOB) to demonstrate that optimization-induced differences can cause wrong predictions. However, ISOB remains input-specific and cannot establish a universal optimization-triggered backdoor. To overcome this limitation, we design the Universal Optimization Backdoor (UOB). The backdoored model stays benign under unoptimized execution but activates when inference optimization is enabled. We conduct extensive experiments across seven mainstream open-source LLMs, four tasks, and three optimization backends. UOB reaches up to 100\% attack success while largely preserving clean accuracy. To mitigate this vulnerability, we design three defense methods that reduce the backdoor ASR to at most 0.02 while preserving clean accuracy. These results reveal inference optimization as a new LLM security attack surface and motivate defenses against test-deployment disagreement.

cs.CR↗

arXiv2026

Reliability Testing of Medical Model Performance under Distributed Deployment

Distributed inference has become an indispensable part of deploying medical models under practical latency, memory, and throughput constraints. Although modern frameworks improve serving efficiency through tensor parallelism, mixed precision, kernel fusion, and multi-device communication, they are generally assumed to preserve the behavior observed during centralized HuggingFace evaluation. This assumption creates an evaluation-deployment mismatch: a model may pass offline evaluation but produce a different output after the execution stack changes. To address this mismatch, we propose a testing framework and an improved, distributed-execution-sensitive medical-model benchmark that evaluates the same checkpoint and input under a centralized HuggingFace reference and matched distributed deployments. Extensive experiments across language, vision, and multimodal medical models show that execution changes can produce measurable output disagreements. Across supported visual settings, the test success rate ranges from 0.21 to 0.43 for single-modality models and from 0.32 to 0.98 for multimodal models. The benchmark is aimed at extending medical-model evaluation from capability and security to evaluation-deployment consistency.

eess.IV↗

arXiv2026

Color screening versus thermal decay as the mechanism of $Υ$ suppression in high energy nuclear collisions

To clearly identify the mechanism behind the suppression of heavy quarkonium in relativistic heavy-ion collisions, we study $Υ$ production at RHIC energies by solving its transport equation driven solely by the suppression rates. By calculating the nuclear modification factor and comparing it with experimental data, we find that the sudden suppression governed by the color-screening temperature cannot simultaneously describe both the ground and excited states of the $Υ$, whereas the continuous suppression induced by thermal decay successfully reproduces all the $Υ$ measurements. This provides strong evidence that inelastic scatterings with thermal partons, rather than color screening, dominate quarkonium suppression in heavy-ion collisions.

hep-ph↗

arXiv2026

Investigating $J/ψ$ spin alignment in heavy-ion collisions within a two-component transport model

We investigate the spin alignment of $J/ψ$ mesons in relativistic heavy-ion collisions within a two-component Boltzmann transport model. Starting from the relativistic spin Boltzmann equation, we derive the spin density matrix element $ρ_{00}$ under a non-relativistic approximation for heavy quarks. To interpret the recent ALICE measurements in Pb+Pb collisions, the observed $ρ_{00}$ is described as a $p_T$-dependent mixture of contributions from primordial production and the coalescence process. At forward rapidity, the $p_T$ dependence of charmonium $ρ_{00}$ is well reproduced by this two-component mechanism: at low $p_T$, charmonium production is dominated by the coalescence of partially polarized charm quarks induced by thermal vorticity; with increasing $p_T$, primordially produced charmonia become dominant, causing $ρ_{00}$ to approach $1/3$. To further test this spin alignment mechanism, we provide predictions for the $J/ψ$ $ρ_{00}$ in the mid-rapidity region, which exhibits a distinct $p_T$ trend due to the kinematic suppression of the thermal vorticity contribution. This study elucidates the underlying mechanism of $J/ψ$ spin alignment and advances our understanding of heavy quarkonium spin dynamics in strongly interacting matter.

nucl-th↗

arXiv2026

Nonlinear Breakdown of Antisymmetric Flexoelectric Coupling

Flexoelectric coupling is conventionally regarded as antisymmetric, such that reversing the strain gradient reverses the polarization direction while preserving its magnitude. Here we report that this antisymmetric coupling breaks down in noncentrosymmetric single crystals driven by large strain gradients, where nonlinearity becomes operative. Harmonic-resolved measurements reveal robust even-order flexoelectric harmonics that are forbidden in antisymmetric systems, as well as the three-wave mixing indicative of symmetry breaking under dual-frequency excitation. These frequency-relevant generations provide evidence for asymmetric flexoelectricity in the nonlinear regime. A reconstructed nonlinear constitutive framework consistently accounts for all observations by incorporating even-order strain-gradient terms. Our findings establish asymmetric flexoelectricity as an intrinsic electromechanical response and uncover a symmetry-breaking mechanism in flexoelectricity.

cond-mat.mtrl-sci↗

arXiv2025

Beyond Playtesting: A Generative Multi-Agent Simulation System for Massively Multiplayer Online Games

Optimizing numerical systems and mechanism design is crucial for enhancing player experience in Massively Multiplayer Online (MMO) games. Traditional optimization approaches rely on large-scale online experiments or parameter tuning over predefined statistical models, which are costly, time-consuming, and may disrupt player experience. Although simplified offline simulation systems are often adopted as alternatives, their limited fidelity prevents agents from accurately mimicking real player reasoning and reactions to interventions. To address these limitations, we propose a generative agent-based MMO simulation system empowered by Large Language Models (LLMs). By applying Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) on large-scale real player behavioral data, we adapt LLMs from general priors to game-specific domains, enabling realistic and interpretable player decision-making. In parallel, a data-driven environment model trained on real gameplay logs reconstructs dynamic in-game systems. Experiments demonstrate strong consistency with real-world player behaviors and plausible causal responses under interventions, providing a reliable, interpretable, and cost-efficient framework for data-driven numerical design optimization.

cs.AI↗

arXiv2025

Impact of dark matter on strange quark stars described by different quark models

Dark matter is hypothesized to interact with ordinary matter solely through gravity and may be present in compact objects such as strange quark stars. We treat strange quark stars admixed with dark matter as two-fluid systems to investigate the potential effects of dark matter on strange quark stars. Quark matter is described by the quasiparticle model and the extended MIT bag model for comparison. Dark matter is treated as asymmetric, self-interacting, and composed of massive fermionic particles. The two-fluid Tolman-Oppenheimer-Volkoff (TOV) equations are employed to solve for specific stellar properties. Our analysis yields relations between central energy density and mass, radius and mass, as well as tidal deformability and mass. The calculated curves generally align with observational data. In particular, we find that the pattern in which fermionic asymmetric dark matter affects the properties of strange quark stars may not be influenced by the equation of state (EOS) of strange quark matter.

astro-ph.HE↗

arXiv2024

Effects of phase transition in hybrid stars from quark-meson coupling hadronic matter to deconfined quark matter

Different types of phase transition from hadron to quark at high density near zero temperature may occur in the inner core of hybrid stars. We investigate the impacts of phase transition types and quark models on properties of hybrid star and quark cores. The quark-meson coupling (QMC) model is used to describe hadronic matter, and the MIT bag model as well as the Nambu-Jona-Lasinio (NJL) model are employed to describe quark matter for comparison. From the mass-radius curves obtained by using equations of state (EOS), we find that EOSs of hadron matter have a decisive influence on the maximum mass of a hybrid star in the first-order phase transition case, while quark matter EOSs have more impacts on results in the crossover transition case. It is also found in the present work that the thermodynamic correction arising from an interpolation scheme considerably stiffens the EOSs. Therefore the crossover type phase transition generally leads to hybrid stars with higher masses. In particular, by using the QMC model and the NJL model to construct crossover EOSs with thermodynamic correction, we discover that the maximum masses of hybrid stars can meet the recent observational constraint on mass from PSR J0952-0607, i.e., $2.35\pm0.17M_\odot$.

nucl-th↗

arXiv2018

An atomistic model for the charge distribution in layered MoS2

We present an atomistic model for predicting the distribution of doping electric charges in layered molybdenum disulfide (MoS$_{2}$). This model mimics the charge around each ion as a net Gaussian-spatially-distributed charge plus an induced dipole, and is able to predict the distribution of doping charges in layered MoS$_{2}$ in a self-consistent scheme. The profiles of doping charges in monolayer MoS$_{2}$ flakes computed by this charge-dipole model are in good agreement with those obtained by density-functional-theory calculations. Using this model, we quantitatively predict the charge enhancement effect in MoS$_{2}$ monolayer nanoribbons, with which strong ionic charge-localization effects are shown.

cond-mat.mes-hall↗