arXiv subjects

Yue Pan

Publications and source records attributed to Yue Pan.

At least 19 recordsLinked to original sources

arXiv2026

Brain-Token Learning: Microstate-Based Tokenization and Multi-Scale Interaction for Long-Horizon EEG Sequence Modeling

Electroencephalography (EEG) provides a non-invasive window into dynamic brain activity, yet modeling long-horizon EEG sequences remains challenging due to their high temporal complexity, substantial variability across subjects, and the lack of biologically meaningful sequence representations. Existing tokenization strategies, such as fixed-window and patch-based representations, discretize EEG signals according to artificial temporal boundaries, which may disrupt intrinsic brain-state dynamics. In this work, we propose Brain-Token Learning, a neuroscience-inspired framework that introduces Brain Tokenization for long-horizon EEG sequence modeling. Instead of partitioning EEG signals into predefined temporal segments, Brain Tokenization represents EEG as sequences of recurrent microstate-derived brain tokens, where each token corresponds to a quasi-stable large-scale brain state with variable temporal duration. Based on these biologically grounded tokens, we further develop a multi-scale token interaction module consisting of Latent State Aggregation and State Transition Modeling to jointly capture global brain-state context and local microstate transitions. We evaluate Brain-Token on five heterogeneous EEG datasets, including the newly collected long-horizon NeuroLong dataset and four affective or clinical EEG datasets (SEED, DEAP, MDD, and NSSI). Extensive experiments demonstrate that Brain-Token consistently outperforms conventional CNN/LSTM architectures, Transformer-based models, and domain adaptation methods across diverse EEG scenarios. Further analysis verifies the effectiveness of microstate-based tokenization and multi-scale interaction for learning robust and interpretable EEG representations. These results establish Brain-Token as a biologically grounded tokenization paradigm for long-horizon EEG sequence modeling.

cs.AI↗

arXiv2026

Optical investigation of the electronic structure of a ferromagnetic Weyl semimetal CeAlSi

To investigate electronic states during the ferromagnetic transition in a magnetic Weyl semimetal CeAlSi, we measured temperature-dependent optical conductivity [$σ_1(ω)$] spectra and compared them with DFT+DMFT band calculations. The $σ_1(ω)$ spectrum did not change significantly across the ferromagnetic ordering temperature ($T_C$), suggesting that the Ce 4f states are almost localized. DFT+DMFT calculations with almost localized Ce 4f states successfully reproduced the spectral shape and the unchanged $σ_1(ω)$ spectra across $T_C$. The dynamic effective mass evaluated from the extended Drude model is very small, which DFT+DMFT calculations also reproduce, but the scattering probability at even lower temperatures suggests ferromagnetic fluctuations. These results suggest that the interaction intensity between the Weyl fermions and Ce 4f states is very weak, as reproduced by DFT+DMFT calculations.

cond-mat.str-el↗

arXiv2026

BlueLM-GUI Technical Report: A Real-Device-Centric Flywheel for Self-Improving Mobile GUI Agents

Mobile GUI agents are shifting from multi-module frameworks to native models trained end-to-end, yet industrial deployment faces three persistent gaps. Sandbox training produces a distribution mismatch with production environments; expensive real-device failures remain underutilized; and fixed benchmarks saturate, losing the power to guide iteration. We present BlueLM-GUI, a 35B-A3B mobile GUI agent built as a real-device-centric flywheel that closes these gaps through three principles. Every Sample Matters: a dual-track pipeline with Heterogeneous Triple-System Consensus evaluation and an Error Correction \& Derivation Module salvages every trajectory into usable supervision. Every Rollout Is Real: a three-stage recipe---continual pre-training, supervised fine-tuning, and agentic reinforcement learning on hundreds of real phones---grounds every rollout in real production environments, so the capability the model learns transfers directly to deployment. Every Query Evolves: a quota-driven benchmark methodology with three orthogonal axes enables precise attribution and allows the benchmark to be systematically upgraded as the model improves. BlueLM-GUI achieves 87.4 on MobileGUI-VBench, surpassing the best closed-source model by 5.1 points, and 84.9 on AndroidWorld, the best result among open-source models and competitive with closed-source models. These results demonstrate that grounding model training and iterative improvement in both real devices and the three Every principles yields strong, robust, and transferable mobile GUI capability.

cs.AI↗

arXiv2026

Asymptotic theory and inference for non-stationary and non-mixing triangular arrays of random fields

We develop a unified asymptotic framework for non-stationary, non-mixing triangular arrays of random fields on multi-dimensional lattices under row-uniform $η$-weak dependence. We establish the preservation of weak dependence under locally Lipschitz transformations and spatially and row-wise heterogeneous Bernoulli shifts, obtaining explicit dependence bounds determined by innovation dependence and functional sensitivity. Under uniform moment conditions, polynomially decaying $η$-dependence coefficients, and a nondegenerate aggregate variance condition, we prove a law of large numbers and a central limit theorem. The scope of the framework is illustrated through several examples that highlight its ability to accommodate non-stationarity without requiring mixing assumptions. As an application, we develop parameter inference procedures for a spatio-temporal model with a dynamic network structure, thereby demonstrating the practical relevance of the proposed asymptotic theory.

math.ST↗

arXiv2026

RAMSES-GPU: Cell-by-Cell Adaptive Mesh Refinement with Magneto-Hydrodynamics and Self-Gravity on Graphics Processing Units

We present the implementation and optimization of the cosmological simulation code RAMSES on Graphics Processing Units (GPUs) using CUDA Fortran. This accelerated version ports the main computational routines, including hydrodynamics, particle dynamics, and self-gravity, to multi-GPU architectures. We detail our strategy for managing cell-by-cell Adaptive Mesh Refinement (AMR) on the GPU, utilizing bucket sort with prefix sums for AMR level sorting, radix sort via the CUB library for Hilbert key ordering, and an fnv64 hash table with linear probing for fast spatial indexing. Portability across diverse hardware architectures is achieved via a dispatcher and C-Fortran wrappers, calling CUDA, HIP, and Metal kernels directly translated from the CUDA Fortran framework. Hydrodynamics updates are executed via a Godunov MUSCL-Hancock HLLC Riemann solver managed through a three-tier shared-memory kernel architecture (named rock, paper, and scissor). Particle mass deposition uses Cloud-in-Cell (CIC) interpolation optimized with atomic additions or prefix sums, combined with a kick-drift-kick time integration pusher. Self-gravity is handled via a Multigrid (MG) Poisson solver performing hierarchical V-cycles on individual levels. Performance benchmarks conducted on NVIDIA A100 and H200 GPUs demonstrate substantial accelerations compared to multi-core CPUs, yielding 10x up to a 100x speedup for standard test problems such as the Sedov blast wave, molecular core collapse, and cosmological simulations. Finally, we briefly discuss additional accelerated physics modules, including equilibrium cooling, polytropic equations of state, ideal and non-ideal magneto-hydrodynamics (MHD), and stellar feedback.

astro-ph.IM↗

arXiv2026

Detecting Vulnerability-Inducing Commits via Multi-Stage Reasoning with LLM-Based Agents

Detecting vulnerability-inducing commits (VICs) at submission time is critical for improving the security and reliability of software systems. However, this task is highly challenging because it requires reasoning about the semantic impact of code changes from heterogeneous information sources, including code diffs, commit messages, and the surrounding contextual code. Existing approaches often struggle to fully capture these complex interactions, resulting in limited detection performance. In this paper, we propose VIC-RAGENT, an LLM-based multi-agent framework for effective and explainable vulnerability detection. VIC-RAGENT leverages multiple specialized agents to provide complementary perspectives, including structural analysis, intent understanding, and vulnerability inspection. To further improve detection reliability, the framework employs a multi-stage reasoning process that progressively refines candidate vulnerabilities through preliminary inspection, reanalysis, and a final decision stage. Experimental results on a real-world dataset across multiple LLMs demonstrate that VIC-RAGENT consistently outperforms baselines, including Direct, CoT, and CodeAgent. Compared to the strongest baseline, VIC-RAGENT achieves 1.2-1.7x higher F1-scores across different models. Overall, VIC-RAGENT offers a robust, explainable, and practical solution for detecting VICs in modern software development workflows.

cs.SE↗

arXiv2026

FORGE: Research-Trajectory Hijacking Attacks on Deep Research Agents

Deep research agents decompose open-ended queries into subtasks, retrieve web evidence over multiple rounds, and synthesize long-form reports. This workflow creates a planning-layer poisoning surface: adversarial documents that enter the retrieval pool can steer follow-up questions and turn a local injection into report-level contamination. We present FORGE (Fabricated Orchestrated Reasoning chain for aGent Exploitation), a two-level attack that combines intra-document reasoning fabrication with inter-document chain coordination to hijack subtask planning. We further introduce the PRISM metric, which weights infected report claims by cognitive type, and Root Query Anchoring, a lightweight defense that ties recursive follow-up generation to the root query. Across 25 queries, Network FORGE reaches 26.4% PRISM with five injected documents and exhibits depth migration, in which recursive synthesis shifts poisoned content from overt framing into factual premises. On the 10-query defense subset, RQA (Root Query Anchoring) reduces PRISM from 38.5% to 18.3%.

cs.AI↗

arXiv2026

Quantum entanglement of photon pairs at proton-proton colliders

Diphoton systems, with photon polarizations measurable at low energies, serve as ideal qubits and were first used to demonstrate quantum entanglement. However, due to the current absence of dedicated polarimeters at high-energy colliders, the entanglement properties of diphoton systems remain largely unexplored at the high-energy frontier. Studying quantum entanglement at the high-energy frontier, where particle colliders serve as a natural relativistic laboratory, helps us better understand the quantum nature and seek new physics. In this letter, we propose a novel method to measure the entanglement of diphoton systems at proton-proton colliders. The photon spin analyzing power arises from the Bethe-Heitler process occurring in the tracker, where photons scatter off nuclei to produce electron-positron pairs whose joint angular distribution encodes the polarization of the diphoton system. Simulation results show that, under HL-LHC conditions, the statistical significance of quantum entanglement in the Higgs $\to γγ$ process is $0.007σ$, while measuring the continuum diphoton process $q\bar{q} \to γγ$ alone can reach about $1.5σ$.

hep-ph↗

arXiv2026

Cluster, Route, Escalate: Cascaded Framework for Cost-Aware LLM Serving

Efficient deployment of large language models (LLMs) in production forces a trade-off between accuracy and cost. Operators often default to a single model that is either expensive for easy queries or insufficient for hard ones. To address this challenge, we propose a two-stage cascaded solution. Stage 1 clusters incoming queries and assigns each cluster to its most cost-effective model. The cost budget for this routing process is set by an interpretable hyperparameter, tuned offline. Stage 2 adds a quality estimation (QE) cascade; when an output from Stage 1 is judged low-quality, the query is escalated to a stronger model. This ensures only hard or low-confidence cases reach the expensive models. On the test datasets, the cascaded system retains 97-99% of the strongest model's accuracy while reducing Time Per Output Token (TPOT). It requires only task-correctness labels and adapts to changes in the model pool without manual reconfiguration.

cs.PF↗

arXiv2026

The Pitfall of Scaling Up: Uncovering and Mitigating Popularity Bias Amplification in Scaling Transformer-based Recommenders

We identify a critical pitfall in scaling transformer-based sequential recommenders: while increasing model size improves recommendation accuracy, it simultaneously amplifies popularity bias. This bias drives systems to over-recommend popular items at the expense of niche ones, which not only undermines fairness but also degrades the broader ecosystem by reinforcing the Matthew effect and filter bubbles. Consequently, this bias amplification emerges as a fundamental obstacle to sustainable model scaling. Through comprehensive theoretical and empirical analyses, we uncover the root cause of this amplification. Our findings reveal that as model depth increases, the two core components of the transformer architecture, i.e., attention aggregation and feed-forward projections, synergistically induce severe spectral collapse in model predictions, which directly translates to the amplification of popularity bias. To address this challenge, we propose SPRINT (Scalable Popularity Regularization IN Transformers), which mitigates spectral collapse during scaling by constraining (i) the maximum column-sums of the attention score matrices and (ii) the spectral norms of the feed-forward parameters. Extensive experiments demonstrate that SPRINT significantly improves both accuracy and long-tail fairness. Crucially, it yields more favorable scaling behaviors when expanding model sizes from 0.05M to 0.34B parameters. The code is available at https://github.com/Tiny-Snow/GenRec.

cs.IR↗

arXiv2026

MapSatisfyBench: Benchmarking Satisfaction-Aware Map Agents through Behavior-Grounded Implicit Decision Factors

Large language model agents are increasingly integrated into map services. Since map services are embedded in everyday-life scenarios rather than professional task settings, users often express their needs informally, resulting in underspecified queries with many unspoken needs, namely, implicit decision factors that are critical for user satisfaction. Although clarification is an effective way to mitigate this issue, it increases user burden in daily interaction, and a capable agent should first proactively recover such factors from available information sources. However, evaluating this ability is challenging. The first challenge is to determine which implicit decision factors are suitable for evaluation. A factor is evaluable only if it affects user acceptance and can be recovered from information available to the agent before it responds. Second, user satisfaction cannot be reliably represented by a single reference answer, requiring a benchmark that converts satisfaction-relevant factors into objective and quantifiable evaluation targets. To address these challenges, we propose a restore-identify-filter framework that reconstructs complete user needs from behavior-chain evidence, identifies implicit decision factors, and retains only those supported by pre-query evidence. Building on this methodology, we construct MapSatisfyBench from large-scale, real-world anonymized user data and annotate ground truth from five dimensions and enables full-chain evaluation of satisfaction-aware map agents. Experiments show that current agents generally perform well on explicit task completion, but remain limited in satisfying implicit decision factors and proactively acquiring the evidence needed for satisfaction-aware decisions. These findings establish MapSatisfyBench as a benchmark for shifting map-agent evaluation from task completion toward satisfaction-aware spatial decision making.

cs.AI↗

arXiv2026

Curvature-Guided Geometric Representation for Protein-Ligand Binding Affinity Prediction

Protein-ligand binding affinity (PLA) prediction is critical in drug discovery. Despite the notable advancements in machine learning-based approaches, existing methods struggle to jointly characterize local geometric organization and globally coordinated cross-molecular interactions, limiting their ability to model complex binding mechanisms. Here, we propose RicciBind, a geometric representation framework that integrates curvature-guided hierarchical structure learning with optimal transport (OT)-based cross-domain alignment to model molecular interactions. Specifically, RicciBind leverages Ricci curvature to capture local interaction tightness within molecular structures, enhancing structural awareness and organizing atomic interactions into curvature-aware hierarchical representations. An OT-based cluster matching mechanism then aligns protein and ligand clusters across heterogeneous domains under geometric constraints, enabling globally consistent correspondences and revealing higher-order interaction patterns beyond local neighborhoods. By coupling curvature-guided structure encoding with OT-driven cross-domain alignment, RicciBind effectively models complex interaction semantics and substantially improves both the accuracy and interpretability of binding affinity prediction. Extensive experiments demonstrate that RicciBind achieved superior predictive performance and generalization across PLA benchmarks and virtual screening tasks. Ablation studies further confirmed the essential role of Ricci curvature in enhancing molecular interaction representations.

cs.LG↗

arXiv2026

Experimental Tabletop Petz recovery of a photonic qubit

The quantum information lost in open evolutions cannot be fully recovered, but partial recovery is possible. The Petz recovery map guarantees almost optimal recovery, notably if the chosen reference state is close to the real one. This map has been widely used in theoretical studies, but has been the object of only a handful of experimental realisations, typically under a single fixed noise model. In this work, we describe and implement the Petz recovery map for a versatile class of qubit channels with tunable decoherence and dissipation. The setup we realize is also the first experimental example of ``tabletop reversibility'': for a good range of choices of the reference state, the Petz recovery map can be implemented with the same devices as the forward dissipative evolution, whose effect it is partially undoing. Our results demonstrate that the Petz recovery map can be resource-efficiently realized without requiring complex ancillary resources, providing a feasible pathway for mitigating information loss in quantum systems.

quant-ph↗

arXiv2026

EgoIntrospect: An Egocentric Dataset and Benchmark for User-Centric Internal State Reasoning

Despite extensive efforts on egocentric video datasets and benchmarks, understanding users' internal states, which is crucial for enabling seamless AI assistant experiences, remains largely overlooked. In this work, we introduce EgoIntrospect, the first egocentric dataset captured in user-driven scenarios with self-annotations that explicitly reveal users' interactive intentions with AI assistants. EgoIntrospect was collected using a cross-device setup, providing synchronized video, audio, gaze, motion, and physiological signals. It consists of 180 hours of recordings from 60 subjects, with an average recording duration of 3 hours per subject. Leveraging EgoIntrospect, we formalize a suite of tasks centered on user internal states, including affective experience, interactive intent, and cognitive memory. We further process the annotations to construct benchmarks that evaluate the ability of modern multimodal large language models to reason about users' internal states from egocentric observations. Experiments on our benchmark suggest that existing multimodal large language models struggle to effectively leverage multimodal signals to infer users' subjective internal states. The dataset and annotations will be made publicly available to advance research in egocentric vision and wearable AI assistants. Project page: https://ego-introspect.github.io/

cs.CV↗

arXiv2026

SVOM/VT: Flight Model Verification and Pre-launch Testing

This paper presents pre-launch testing and calibration results for the SVOM/VT (Space-based Variable Objects Monitor, Visible Telescope) Flight Model (FM), validating its performance under simulated space conditions through thermal vacuum cycling, energy concentration analysis, stray light suppression, and CCD/electronics calibrations (gain, noise, quantum efficiency). The results confirm full compliance with design requirements: stray light suppression achieves point-source transmittance $<10^{-7}$ at $30^\circ$ off-axis, thermal control maintains stable CCD temperatures ($-75^\circ$C for the red channel, $-65^\circ$C for the blue channel), and detection sensitivity meets the limiting magnitude of 22.50 (SNR $>$ 3 with 300 seconds exposure). Early in-orbit tests further validate performance, yielding limiting magnitudes of 22.70 (V-band, red) and 22.78 (blue), consistent with pre-launch specifications.

astro-ph.IM↗

arXiv2026

SOLAR-RL: Semi-Online Long-horizon Assignment Reinforcement Learning

As Multimodal Large Language Models (MLLMs) mature, GUI agents are evolving from static interactions to complex navigation. While Reinforcement Learning (RL) has emerged as a promising paradigm for training MLLM agents on dynamic GUI tasks, its effective application faces a dilemma. Standard Offline RL often relies on static step-level data, neglecting global trajectory semantics such as task completion and execution quality. Conversely, Online RL captures the long-term dynamics but suffers from high interaction costs and potential environmental instability. To bridge this gap, we propose SOLAR-RL (Semi-Online Long-horizon Assignment Reinforcement Learning). Instead of relying solely on expensive online interactions, our framework integrates global trajectory insights directly into the offline learning process. Specifically, we reconstruct diverse rollout candidates from static data, detect the first failure point using per-step validity signals, and retroactively assign dense step-level rewards with target-aligned shaping to reflect trajectory-level execution quality, effectively simulating online feedback without interaction costs. Extensive experiments demonstrate that SOLAR-RL significantly improves long-horizon task completion rates and robustness compared to strong baselines, offering a sample-efficient solution for autonomous GUI navigation.

cs.LG↗

arXiv2026

Skill-SD: Skill-Conditioned Self-Distillation for Multi-turn LLM Agents

Reinforcement learning (RL) has been widely used to train LLM agents for multi-turn interactive tasks, but its sample efficiency is severely limited by sparse rewards and long horizons. On-policy self-distillation (OPSD) alleviates this by providing dense token-level supervision from a privileged teacher that has access to ground-truth answers. However, such fixed privileged information cannot capture the diverse valid strategies in agent tasks, and naively combining OPSD with RL often leads to training collapse. To address these limitations, we introduce Skill-SD, a framework that turns the agent's own trajectories into dynamic training-only supervision. Completed trajectories are summarized into compact natural language skills that describe successful behaviors, mistakes, and workflows. These skills serve as dynamic privileged information conditioning only the teacher, while the student always acts under the plain task prompt and learns to internalize the guidance through distillation. To stabilize the training, we derive an importance-weighted reverse-KL loss to provide gradient-correct token-level distillation, and dynamically synchronize the teacher with the improving student. Experimental results on agentic benchmarks demonstrate that Skill-SD substantially outperforms the standard RL baseline, improving both vanilla GRPO (+14.0%/+10.9% on AppWorld/Sokoban) and vanilla OPD (+42.1%/+40.6%). Project page: https://k1xe.github.io/skill-sd/

cs.LG↗

arXiv2026

An Intention-driven Lane Change Framework Considering Heterogeneous Dynamic Cooperation in Mixed-traffic Environment

In mixed-traffic environments, autonomous vehicles (AVs) must interact with heterogeneous human-driven vehicles (HVs) whose intentions and driving styles vary across individuals and scenarios. Such variability introduces uncertainty into lane change interactions, where safety and efficiency critically depend on accurately anticipating surrounding drivers' cooperative responses. Existing methods often oversimplify these interactions by assuming uniform or fixed behavioral patterns. To address this limitation, we propose an intention-driven lane change framework that integrates driving-style recognition with cooperation-aware decision-making and motion-planning. A deep learning-based classifier identifies distinct human driving styles in real time. We then introduce a dual-perspective cooperation score composed of intrinsic style-dependent tendencies and interactive dynamic components, enabling interpretable and adaptive intention prediction and quantitative inference. A decision-making module combines behavior cloning (BC) and inverse reinforcement learning (IRL) to determine lane change feasibility. Later, a coordinated motion-planning architecture integrating IRL-based intention inference with model predictive control (MPC) is established to generate collision-free and socially compliant trajectories. Experiments on the NGSIM dataset show that the proposed decision-making model outperforms representative rule-based and learning-based baselines, achieving 96.98% accuracy in lane change classification. Motion-planning evaluations further demonstrate improved maneuver success and execution stability in mixed-traffic environments. These results validate the effectiveness of structured cooperation modeling for intention-driven autonomous lane changes.

cs.RO↗