arXiv subjects

Like Liu

Publications and source records attributed to Like Liu.

8 recordsLinked to original sources

arXiv2026

Effects of femtoscopic correlations on spin-spin correlation measurements

Hyperon spin correlations serve as sensitive probes of spin dynamics in high-energy collisions, yet their extraction from weak-decay angular distributions can be contaminated by femtoscopic effects due to quantum statistics and final-state interactions. In this work, we quantitatively assess this contamination for $\Lambda\Lambda$ and $\Lambda\bar{\Lambda}$ pairs using the AMPT model combined with spin-dependent weights from the Lednick\'y--Lyuboshits formalism. Because the singlet and triplet spin configurations contribute differently to the decay-angle distribution, femtoscopic weighting induces an apparent angular modulation, creating a fake correlation signal even when no intrinsic spin correlation is present. We find that the induced bias can become substantial in the low-$q_{\mathrm{inv}}$ region, where the combined femtoscopic effect reaches a magnitude comparable to that of the preliminary CMS measurements. Our results establish a framework for evaluating such systematics, highlighting that femtoscopic corrections must be carefully considered in future differential analyses that emphasize the low-relative-momentum region.

nucl-th↗

arXiv2026

Knowing the Self, Understanding the World: A Dual-Cognition Benchmark for UAV Spatio-temporal Reasoning with MLLMs

Multimodal large language models have achieved strong performance across diverse vision-language tasks, yet their capabilities in UAV scenarios remain insufficiently explored. Recent UAV-oriented benchmarks have begun to evaluate MLLMs in aerial scenarios, but they typically focus on scene understanding, event recognition, or navigation completion, rather than jointly assessing the dual-cognition capability required for UAV agents: reasoning about both the UAV's own state and the external environment in multiview spatio-temporal contexts. To address this gap, we present UAV-DualCog, a benchmark for aerial multiview spatio-temporal reasoning built on this dual-cognition perspective. UAV-DualCog includes both image and video tasks to jointly evaluate self-state and environment-state reasoning, while requiring spatial or temporal grounding beyond discrete answer prediction. We also develop an automated pipeline that constructs data from scene-level semantic point clouds, yielding a scalable benchmark with diverse scenes, hundreds of landmarks, and thousands of QA samples. Extensive evaluations show that current MLLMs remain far from reliable in UAV dual cognition. Self-state reasoning, viewpoint transformation, precise spatial grounding, and temporal interval localization are persistent bottlenecks, and additional validation with thinking/frontier models and a human baseline confirms that the benchmark is understandable to humans but challenging for existing models. We further construct UAV-DualCog-Train from disjoint scenes and show through a lightweight optimization probe that it provides useful structured supervision, suggesting its value not only as an evaluation benchmark but also as a data resource for advancing MLLM-based UAV agents. Project website and supplementary materials: https://uav-dualcog.lozumi.com

cs.CV↗

arXiv2026

FineCog-Nav: Integrating Fine-grained Cognitive Modules for Zero-shot Multimodal UAV Navigation

UAV vision-language navigation (VLN) requires an agent to navigate complex 3D environments from an egocentric perspective while following ambiguous multi-step instructions over long horizons. Existing zero-shot methods remain limited, as they often rely on large base models, generic prompts, and loosely coordinated modules. In this work, we propose FineCog-Nav, a top-down framework inspired by human cognition that organizes navigation into fine-grained modules for language processing, perception, attention, memory, imagination, reasoning, and decision-making. Each module is driven by a moderate-sized foundation model with role-specific prompts and structured input-output protocols, enabling effective collaboration and improved interpretability. To support fine-grained evaluation, we construct AerialVLN-Fine, a curated benchmark of 300 trajectories derived from AerialVLN, with sentence-level instruction-trajectory alignment and refined instructions containing explicit visual endpoints and landmark references. Experiments show that FineCog-Nav consistently outperforms zero-shot baselines in instruction adherence, long-horizon planning, and generalization to unseen environments. These results suggest the effectiveness of fine-grained cognitive modularization for zero-shot aerial navigation. Project page: https://smartdianlab.github.io/projects-FineCogNav.

cs.CV↗

arXiv2026

Improved Pion-Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation

Accurate identification of charged pions and kaons is essential for precision measurements in relativistic heavy-ion collisions, but becomes increasingly challenging at intermediate and high transverse momentum due to the overlap between time-of-flight mass-square ($m^{2}$) and ionization energy loss ($n\sigma$) distributions. In this work, we present a two-dimensional shift and rotation method that exploits the correlated information between $m^{2}$ and $n\sigma$ to enhance particle identification performance. The method is validated using Au+Au collision events generated with the AMPT model, where detector response effects are incorporated through a data-driven smearing procedure tuned to reproduce the particle identification performance of the STAR experiment. The reconstructed pion and kaon transverse momentum distributions show excellent agreement with the AMPT input, maintaining a purity exceeding 98\% at high $p_T$ and extend the reliable identification range up to $p_T \approx$ 3 GeV/$c$. The extracted elliptic flow $v_2$ remains consistent with the input over the extended $p_T$ range, demonstrating that the proposed method provides a robust framework for high precision identified hadron measurements.

physics.ins-det↗

arXiv2026

Learning to Generate Secure Code via Token-Level Rewards

Large language models (LLMs) have demonstrated strong capabilities in code generation, yet they remain prone to producing security vulnerabilities. Existing approaches commonly suffer from two key limitations: the scarcity of high-quality security data and coarse-grained reinforcement learning reward signals. To address these challenges, we propose Vul2Safe, a new secure code generation framework that leverages LLM self-reflection to construct high-confidence repair pairs from real-world vulnerabilities, and further generates diverse implicit prompts to build the PrimeVul+ dataset. Meanwhile, we introduce SRCode, a novel training framework that pioneers the use of token-level rewards in reinforcement learning for code security, which enables the model to continuously attend to and reinforce critical fine-grained security patterns during training. Compared with traditional instance-level reward schemes, our approach allows for more precise optimization of local security implementations. Extensive experiments show that PrimeVul+ and SRCode substantially reduce security vulnerabilities in generated code while improving overall code quality across multiple benchmarks.

cs.CR↗

arXiv2025

FineTec: Fine-Grained Action Recognition Under Temporal Corruption via Skeleton Decomposition and Sequence Completion

Recognizing fine-grained actions from temporally corrupted skeleton sequences remains a significant challenge, particularly in real-world scenarios where online pose estimation often yields substantial missing data. Existing methods often struggle to accurately recover temporal dynamics and fine-grained spatial structures, resulting in the loss of subtle motion cues crucial for distinguishing similar actions. To address this, we propose FineTec, a unified framework for Fine-grained action recognition under Temporal Corruption. FineTec first restores a base skeleton sequence from corrupted input using context-aware completion with diverse temporal masking. Next, a skeleton-based spatial decomposition module partitions the skeleton into five semantic regions, further divides them into dynamic and static subgroups based on motion variance, and generates two augmented skeleton sequences via targeted perturbation. These, along with the base sequence, are then processed by a physics-driven estimation module, which utilizes Lagrangian dynamics to estimate joint accelerations. Finally, both the fused skeleton position sequence and the fused acceleration sequence are jointly fed into a GCN-based action recognition head. Extensive experiments on both coarse-grained (NTU-60, NTU-120) and fine-grained (Gym99, Gym288) benchmarks show that FineTec significantly outperforms previous methods under various levels of temporal corruption. Specifically, FineTec achieves top-1 accuracies of 89.1% and 78.1% on the challenging Gym99-severe and Gym288-severe settings, respectively, demonstrating its robustness and generalizability. Code and datasets could be found at https://smartdianlab.github.io/projects-FineTec/.

cs.CV↗

arXiv2024

Elliptic and quadrangular flow of protons in the high baryon density region

The collective flow provides valuable insights into the anisotropic expansion of particles produced in heavy-ion collisions and is sensitive to the equation of the state of nuclear matter in high-baryon-density regions. In this paper, we use the hadronic transport model SMASH to investigate the elliptic flow ($v_2$), quadrangular flow ($v_4$), and their ratio ($v_{4}/v_{2}^{2}$) in Au+Au collisions at high baryon density. Our results show that the inclusion of baryonic mean-field potential in the model successfully reproduces experimental data from the HADES experiment, indicating that baryonic interactions play an important role in shaping anisotropic flow. In addition to comparing the transverse momentum ($p_T$), rapidity, and centrality dependence of $v_{4}/v_{2}^{2}$ between HADES data and model calculations, we also explore its time evolution and energy dependence across $\sqrt{s_{NN}} =$ 2.4 to 4.5 GeV. While the ratio $v_{4}/v_{2}^{2}$ for high-$p_{T}$ particles approaches 0.5, which aligns with expectations from hydrodynamic behavior, we emphasize that this result primarily reflects agreement with the HADES measurements rather than a definitive indication of ideal fluid behavior. These findings contribute to understanding the early-stage dynamics in heavy-ion collisions at high baryon density.

nucl-th↗

arXiv2022

Elliptic and triangular flow of (multi-)strange hadrons and $\phi$ mesons in BES-II energies at STAR

In these proceedings, we present the measurements of elliptic ($v_2$) and triangular ($v_3$) flow of (multi-)strange hadrons and $\phi$ mesons in 19.6 and 14.6 GeV Au+Au collisions from STAR. The number of constituent quark (NCQ) scaling of $v_2$ and $v_3$ holds well at $\sqrt{ s_{\mathrm{ NN }} } $ = 19.6 GeV, which indicates the collective flow is built up in the partonic stage. At these energies, the anti-particles show better NCQ scaling than the particles for both $v_2$ and $v_3$, which may be caused by the different contributions from the produced and transported quarks.

nucl-ex↗