Searcharxiv⌕ Search

arXiv subjects

Xingyu Zhang

Publications and source records attributed to Xingyu Zhang.

At least 19 recordsLinked to original sources
arXiv2026

Variational Framework for Quantum Molecular Reaction Dynamics

This work describes a geometric framework on molecular reaction dynamics based on the variational principle, where the Schr{ö}dinger equation must be solved to ``see'' how a reaction occurs. First, the mathematical preliminaries are given by discussing the principle of least action and the mountain pass theorem. Second, we discuss the physical preliminaries, including the principle of equivalence for deriving the kinetic energy operator (KEO) and artificial intelligence (AI) techniques to build the potential energy surface (PES) in general spacetime. Moreover, we simplified electromagnetic interactions in curved spacetime within the molecular system and consequently, we are able to construct the nuclear Hamiltonian in nonzero curvature spacetime. This indicates possibility to introduce gauge fields through the curvature, such as additional term in the nuclear KEO near a conical intersection. Third, the single-particle approximation provides a powful ansatz to solve the Schr{ö}dinger equation by variational principle. Thus, one can formulate the variational approaches for either electronic structure or quantum dynamics. In this work, based on previous discussions ({\it Phys. Chem. Chem. Phys.} {\bf 27} (2025), 20397) we unified them by a geometric description, where the geometric phase is naturally introduced. Finally, due to optimization characteristic of the present theory, further discussions on the present theory from optimization insight are also given, including two postulates, generative AI techniques, role of perturbation, and Markov process in optimization.

physics.chem-ph↗
arXiv2026

Uncertainty-Driven Self-Refinement of Adaptive Kolmogorov-Arnold Networks for Fitting Nonadiabatic Potential Energy Surface

In this work, we propose an uncertainty-driven self-refinement learning method using adaptive Chebyshev-polynomial-based Kolmogorov-Arnold networks (ChebyKANs) to build potential energy surface (PES). Unlike conventional approaches, our method employs a main-model-plus-residual-model architecture for automatic sampling based on a hybrid criterion, where sampling and fitting are decoupled. To benchmark our method, the non-adiabatic two-state PES of the H + H$_2$ system is constructed using a database computed at the multireference configuration interaction (MRCI) level with a rather large active space. Our calculations show that the proposed architectural design makes the efficiency gain, and that KAN offers enhanced interpretability over conventional neural network (NN) fitting methods through its mathematically well-understood basis functions. Furthermore, the present calculations accurately reproduce all important conformations on the PES, including the transition state and the conical intersection. Then, the multi-configurational time-dependent Hartree (MCTDH) method reproduces the non-adiabatic dynamics with results in close agreement with previously reported findings.

physics.chem-ph↗
arXiv2026

Searching for Dense Perturbers in Cold Dark Matter Simulations

Recently, several low-mass, dense perturbers have been inferred from astrophysical observations of gravitational lensing, stellar streams, and satellite galaxies. If interpreted as dark matter subhalos, their inner densities exceed the predictions of the standard cold dark matter (CDM) model. We systematically investigate this tension using a CDM-only zoom-in simulation of a Milky Way analog. We analyze more than 5000 subhalos with masses down to $10^6\,M_\odot$ and reconstruct their inner density profiles using a tidal truncation model. The simulated subhalos are then compared with the observed compact perturbers in terms of their density profiles, enclosed masses, concentrations, and projected cylindrical masses. None of the simulated CDM subhalos is sufficiently dense to reproduce the inferred mass distributions of the observed perturbers. We further examine the redshift evolution of subhalo concentrations and derive a fitting function to characterize this evolution. Our findings motivate further investigations into exceptionally rare formation histories that could produce such dense perturbers, as well as the possibility of physics beyond the standard CDM model.

astro-ph.CO↗
arXiv2026

Cross-Embodiment Causal Transfer from Human Videos to Robot Manipulation

Scaling robot manipulation models requires large amounts of interaction data, yet collecting diverse real-robot experience remains costly. Human manipulation videos, in contrast, provide abundant and diverse task experience that could potentially benefit robot learning. However, differences in morphology, action spaces, and control mechanisms make human motions unsuitable as direct supervision for robot actions, raising a scientific question: what knowledge in human videos should be transferred across embodiments to robot manipulation? To address this question, we analyze subtask transfer through the lens of causal beams, focusing on which contextual conditions and interaction dependencies should be retained to support local task effects. Based on this perspective, we propose Causal-Beam-Guided Subtask Transfer (CBST). CBST learns manipulation-related representations from human videos and grounds them in task-level semantic variables using robot demonstrations. Within a predefined manipulation model, causal-beam constraints guide contextual information selection by requiring retained assignments to sustain local mechanism outputs while preserving nontrivial responsiveness to retained inputs. The selected semantic information conditions a world-action model, allowing the robot to generate actions from its own observations without requiring human-robot motion correspondence. Experiments on robot manipulation benchmarks show improvements in task success over the corresponding world-action model baselines.

cs.RO↗
arXiv2026

Stationarity equations for fractional affine area measures: ellipsoid rigidity and polytopal solutions

Xi and Zhao established a first variation formula for affine inequalities involving radial mean bodies, and asked two geometric questions about the resulting stationarity equation: whether the bodies already known to satisfy it are the only ones, and whether its smooth solutions are exactly the ellipsoids. We answer both. Every strictly convex solution of the equation with $q>-1$, $q\ne0,n$ is an ellipsoid, which answers the second question affirmatively while replacing smoothness and everywhere positive Gauss curvature by strict convexity. Every full-dimensional facet-transitive polytope satisfies the equation, and so does the polar of every vertex-transitive polytope with respect to its volume centroid; the rhombic dodecahedron, the thirteen Catalan solids and the bipyramids over regular polygons other than the octahedron provide infinitely many affinely inequivalent solutions that are not affine images of regular polytopes or balls, so the answer to the first question is negative. The rigidity proof compares overlap integrals along a continuous Steiner path: stationarity forces an increase of order $o(t)$, whereas a nonaffine chord midpoint function forces an increase of order $t$. For negative parameters the required comparison rests on a minimization formula over bounded measurable auxiliary sets.

math.FA↗
arXiv2026

CAG-WAM: Causal Action-Gain World Action Models for Robot Manipulation

World Action Models (WAMs) combine action generation with future prediction, allowing predicted futures to support action evaluation and robot decision making. However, the future associated with a candidate action may contain action-induced and environment-driven changes. Directly evaluating the predicted future may not fully characterize the future change associated with the candidate action. For example, a candidate action that worsens a favorable environmental future may still yield a favorable outcome, while one that improves an unfavorable environmental future may still yield an unfavorable outcome. To analyze this issue, we formulate a Structural Causal Model (SCM) for action evaluation in WAMs. The SCM formalizes how action-induced and environment-driven changes jointly shape future evaluation, and reveals a backdoor path through which environmental evolution can also affect the future evaluation associated with a candidate action. Intervention and backdoor adjustment account for this path, providing a causal basis for characterizing the future change associated with each candidate action. Based on this analysis, we propose Causal Action-Gain World Action Model (CAG-WAM). For each candidate action, CAG-WAM predicts an action-conditioned future and an environmental future under the same current observation, and defines their evaluation difference as causal action gain (CAG) for characterizing the future change associated with the candidate action. During training, we incorporate CAG into Group Relative Policy Optimization (GRPO) to guide robot policy learning. During inference, we evaluate the initial action with CAG and, when needed, revise it using predicted future information. Experiments on multiple benchmarks support the effectiveness of CAG-WAM.

cs.RO↗
arXiv2026

SparseWAM: Driving World-Action Models are Worth 16 Register Tokens

World Action Models (WAMs) transfer scene-dynamics priors to planning by jointly forecasting future scenes and ego trajectories. However, WAMs built on video generators must generate dense future videos at inference, incurring prohibitive latency, while efficient variants discard the world branch at deployment and thus forgo test-time imagination. Even when dense video latents are replaced with patch-level DINO or JEPA features, these representations still involve long token sequences and encode decision-irrelevant information. We therefore propose SparseWAM, which forecasts in a sparse register space of only 16 tokens per frame, compressed from frozen DINOv2 features and shaped by joint planning-and-scoring pretraining to concentrate decision-relevant information. Within this space, SparseWAM employs a modality-decoupled Mixture-of-Transformers with a causal attention mask that lets action tokens condition on imagined future registers, retaining explicit imagine-then-act reasoning at deployment. Its compactness further enables low-cost multimodal sampling. As a result, our SparseWAM achieves 92.6 PDMS on NAVSIM v1, 90.1 EPDMS on NAVSIM v2, and 39.6 on NavHard at only 164 ms latency, roughly 1/5 that of DriveVA, demonstrating the efficiency of the sparse register space. Code will be released upon acceptance.

cs.CV↗
arXiv2026

Energy-Efficient Non-Volatile Photonic Switching via Composition-Engineered Sn-Doped GST

Integrated phase-change photonics allows non-volatile optical switching with no static power draw, but the electrical energy needed to reversibly switch the material's state limits practical scaling. Efforts to improve energy efficiency have mostly focused on device optimization, such as redesigning the integrated microheater or waveguide. Here, we take a different, materials-centered approach by alloying Sn to Ge$_2$Sb$_2$Te$_5$ (GST) without altering the device structure or CMOS compatibility. We explore Sn concentrations ranging from 0 at.% to 20 at.% on waveguide-integrated PN and PIN heaters and observe a reduction in both the amorphization and crystallization energies in the 5--10 at.% range. At higher Sn content this trend reverses and the switching energy increases, which is suggestive of phase segregation when combined with TEM compositional analysis. A 10 at.% Sn device continues to switch over 1000 cycles, with observable degradation in the extinction ratio from 4.5 $\pm$ 0.3 dB to 2.4 $\pm$ 0.3 dB. Post-cycling imaging shows PCM migration and void formation which we believe can be further improved by enhanced encapsulation. Overall, Sn alloying offers a route to lower programming energy in integrated phase-change photonics without negatively affecting the switching speed relative to un-doped GST.

cond-mat.mtrl-sci↗
arXiv2026

SIDM and CDM interpretations of the million-solar-mass lensing perturber JVAS B1938+666-$\mathcal{V}$

A $10^6\,M_\odot$ object has recently been inferred from gravitational imaging of the strong-lensing system JVAS B1938+666, exhibiting an unusually dense inner region embedded within an extended envelope, far exceeding expectations for cold dark matter (CDM) halos. Using gravothermal fluid simulations, we show that such a structure arises naturally in self-interacting dark matter (SIDM) halos evolving into a deep core-collapse phase, where a secondary dense central core forms within an extended profile. The resulting density structure closely matches the inferred properties of the lensing object. We also demonstrate that a similar profile could be reproduced in CDM in the presence of an intermediate-mass black hole, but this requires an early-forming progenitor that subsequently loses $5$ orders of magnitude in mass through tidal stripping by the lens galaxy. Whether such a scenario can be realized in realistic cosmological environments remains an open question.

astro-ph.GA↗
arXiv2026

ForeDrive: Foresight-Guided End-to-End Autonomous Driving with a Planning-Relevant Latent World Model

Existing latent world models are typically optimized for future predictability, yet the resulting representations are not necessarily useful for planning in autonomous driving. Predictions are commonly used for pretraining or auxiliary supervision rather than as direct conditioning signals for trajectory generation. We propose ForeDrive, which learns a planning-relevant latent representation and couples it asymmetrically to a Diffusion Transformer (DiT) planner. The planner consumes multi-horizon latent future representations learned with a JEPA-style world model; planning gradients update the shared online encoder, while stop-gradient routing trains the latent predictor with forecasting losses only. Because predicted futures have varying reliability across horizons and BEV trajectories are misaligned with image tokens, we use gated visual fusion, future-status injection, and Trajectory-Adaptive Bias (TAB) to inject future latents as guidance without overriding the current observation. Trained with pure imitation learning and using only the current front-view image as visual input at inference, ForeDrive attains 89.9 PDMS on NAVSIM v1 and 90.0 one-stage EPDMS on NAVSIM v2, without reinforcement learning or an external trajectory scorer.

cs.CV↗
arXiv2026

Reflex-Informed Neuromuscular Reinforcement Learning for Muscle-Driven Locomotion

Muscle-driven locomotion provides a physically grounded approach to generating realistic human movement. However, achieving both physiological plausibility and adaptability to changes in musculoskeletal capacity and external disturbances remains a fundamental challenge. To address this limitation, we propose a Reflex-Informed Neuromuscular Reinforcement Learning framework for muscle-driven locomotion. Within this framework, a fixed phase-dependent reflex controller serves as the underlying neuromuscular control mechanism, while the reinforcement learning policy produces four biomechanically meaningful residual parameters to modulate key reflex gains and thresholds associated with hip swing, knee support, and ankle propulsion according to the current state. Experimental results demonstrate that the proposed framework generates physiologically plausible locomotion with improved kinematic accuracy and dynamic consistency, as well as better bilateral symmetry and stride-to-stride consistency under nominal walking conditions. The learned policy remains robust under muscle weakness and external perturbations without retraining.

cs.RO↗
arXiv2026

Implementation Possibility of Quantum Simulation for Quantum Molecular Dynamics

In this work, we explore the implementation possibility of quantum simulation for quantum molecular dynamics, in particular for reaction dynamics, though several implementations have already reported through quantum-classical mixed simulations ({\it Acc. Chem. Res.} {\bf 54} (2021), 4229 and {\it J. Phys. Chem. Lett.} {\bf xx} (2026), XXXX). To analyze this aspect, we examine (1) the conjugacy relation between quantum simulator and the target molecular system, (2) the wave function correspondence in quantum algorithm and classical algorithm for multi-dimensional dynamics, (3) problems arisen from real-valued classical algorithms, and finally (4) geometric phase arisen from the separation among the degrees of freedom (DOFs). As is well known, the aforementioned first and second points play fundamental roles in quantum simulation of quantum many-body systems, and the third and fourth points are theoretical issues that might introduce problems in classical and quantum computing. In this work, we mainly focus on the third and fourth points by analysis of the first two points by reviewing previously reported quantum-classical mixed implementations of quantum simulation. We also consider gauge freedom in high-dimensional quantum molecular dynamics that has been introduced recently, and then discuss possibility of advantages and disadvantages of quantum simulation for molecular reaction dynamics.

physics.chem-ph↗
arXiv2026

GUPO: Gradient Uncertainty-aware Policy Optimization for Post-Training Large Language Models

Group Relative Policy Optimization (GRPO) has become a widely used approach for post-training Large Language Models (LLMs) for reasoning. In GRPO, the group gradients induced by different queries within the same mini-batch are directly averaged to form the policy update. However, these group gradients can point in conflicting directions. Our empirical analysis suggests that group-gradient conflicts tend to be associated with less effective policy updates, motivating the need for a reliable aggregated update direction under such conflicts. Standard GRPO aggregation treats the realized group gradients as deterministic contributions and does not account for differences in their reliability during aggregation. To address this issue, we propose Gradient Uncertainty-Aware Policy Optimization (GUPO), which models each group gradient as a random variable under a Bayesian formulation and estimates its probability distribution. GUPO then derives gradient uncertainty using a Dirichlet-based formulation and uses it to calibrate the contribution of each group gradient during aggregation. Extensive experiments on multiple benchmarks demonstrate the effectiveness of GUPO.

cs.LG↗
arXiv2026

Closing the Loop: A Control-Theoretic Framework for Provably Stable Time Series Forecasting with LLMs

Large Language Models (LLMs) have recently shown exceptional potential in time series forecasting (TSF), leveraging their inherent sequential reasoning capabilities to model complex temporal dynamics. Existing approaches typically employ an autoregressive generation strategy to adapt LLMs for TSF. However, we identify a theoretical flaw in this paradigm: during inference, the model operates in an open-loop manner, recursively consuming its own generated outputs. This leads to error accumulation, where minor early deviations cascade into significant rollout drift over long horizons. In this paper, we reformulate autoregressive forecasting through the lens of control theory, proposing Feedback-driven LLM (F-LLM), a novel closed-loop framework. Unlike standard methods that passively propagate errors, F-LLM actively stabilizes the trajectory via a learnable residual estimator functioning as a system observer. Furthermore, we provide a mathematical proof that, under explicit contraction assumptions, this closed-loop mechanism guarantees a uniformly bounded step-wise error sequence within the local surrogate dynamics. Extensive experiments demonstrate that F-LLM significantly mitigates error propagation, achieving good performance on time series benchmarks. Our code is publicly available at https://github.com/Zh-XY22/F-LLM.

cs.LG↗
arXiv2026

Ground-Truth Neighborhood Regularization for Reinforcement Learning Post-Training of Time Series Foundation Models

Time series forecasting (TSF) plays an important role in a wide range of real-world applications. Recently, time series foundation models (TSFMs), pretrained on large-scale datasets, have demonstrated strong generalization capabilities and emerged as an important paradigm for TSF. Reinforcement learning (RL) post-training has consequently attracted growing attention as a means of further improving their performance on downstream tasks. However, we find that, in certain forecast regions, RL post-training may gradually shift the output distributions of TSFMs away from the ground truth, thereby limiting their performance. We refer to this phenomenon as \textbf{suboptimal collapse}. Our analysis suggests that difficulty in initially sampling high-quality trajectories near the ground truth is an important contributing factor to suboptimal collapse. To address this issue, we propose Ground-Truth Neighborhood Regularization (GTN-R) for RL post-training of TSFMs. GTN-R uses the ground truth as a reference for locating high-quality regions and guides the model's probability mass toward the ground-truth neighborhood. This increases the probability of sampling high-quality trajectories, mitigates suboptimal collapse, and improves performance. Moreover, GTN-R can be flexibly integrated into various RL methods for TSFMs. Extensive experiments show its effectiveness.

cs.LG↗
arXiv2026

Cold Dark Matter and Self-Interacting Dark Matter Interpretations of Cloud-9

Recently, the Five-hundred-meter Aperture Spherical Telescope discovered a gas-rich hydrogen cloud near M94 in the $21\,{\rm cm}$ band. Lacking an optical counterpart, this object, dubbed Cloud-9, has been identified as a compelling Reionization Limited \textsc{Hi} Cloud (RELHIC). RELHICs provide exceptionally clean laboratories for probing dark matter, free from the baryonic complexities associated with star formation and feedback. We show that the observed hydrogen column density profile of Cloud-9 is consistent with a gas cloud embedded in either a cuspy halo predicted by the standard cold dark matter (CDM) model or a cored halo produced by self-interacting dark matter (SIDM). In both cases, the halo must have an unusually diffuse central density. The best-fitting CDM halo lies around $7σ$ below the cosmological concentration--mass relation, whereas SIDM core-forming halos reduce the tension to only around $3σ$. We further identify Cloud-9 analogs in the Concerto suite of cosmological zoom-in simulations with velocity-dependent SIDM, demonstrating that RELHICs provide a promising new probe of dark matter self-interactions.

astro-ph.GA↗
arXiv2026

TacPrint: A Wearable Fingertip Tactile Sensor for Human-to-Robot Contact Reproduction

Human-centric data collection is emerging as a significant paradigm for robot skill acquisition, but seamlessly integrating low-cost, scalable tactile sensing systems that capture fine-grained fingertip interactions without compromising natural operation remains a key challenge. This reduces the reliability of human-to-robot transfer in contact-rich tasks. In this work, we present TacPrint, a wearable fingertip tactile sensor, where protrusions on the inner surface of the silicone skin are aligned one-to-one with 24 capacitive taxels to enable localized capacitive responses. A real-to-sim-to-real pipeline estimates a 35 $\times$ 26 contact-depth map from 24-channel capacitive signals. Against simulation-generated labels, the model achieved a contact-region RMSE of 0.223 $\pm$ 0.161 mm, a weighted-centroid error of 1.213 $\pm$ 2.379 pixels, and an IoU of 0.829 $\pm$ 0.169. With measured capacitive inputs, the network-predicted depth evaluated at the guide-calibrated contact center showed a mean absolute error of 0.085 $\pm$ 0.057 mm across all 40 controlled trials, while the mean contact-position error was 0.250 $\pm$ 0.208 mm across the 37 trials whose reference contact regions were not truncated by the sensing boundary. In human-to-robot replay, tactile-guided compensation increased grasping and wiping success rates from 0% to 91.67% and 90%, respectively. In closed-loop grasping, dense-depth feedback achieved success rates of 87.5% over all tested positions and 85% under edge-contact conditions, compared with 67.5% and 45% for raw-taxel feedback.

cs.RO↗
arXiv2026

Nonreciprocal Relaxation Acceleration

Driven by recent discoveries regarding the quantum Mpemba effect, the anomalous relaxation dynamics of open quantum systems have garnered significant attention. While expediting thermalization to equilibrium has been extensively studied, dynamically accelerating the convergence toward nonequilibrium steady states remains a formidable challenge. In this article, we find a transient engineered nonreciprocal dissipative channel can provide a shortcut that accelerates convergence to the target reciprocal nonequilibrium steady state for the considered two-mode model and initial states. Using interacting bosonic modes, we demonstrate that the temporal activation of a nonreciprocal channel efficiently suppresses prolonged inter-mode energy oscillations, enforcing a rapid, unidirectional thermal dump into the environment. Counterintuitively, we find that this relaxation speedup is robust and independent of the direction of the nonreciprocity. Our results provide a powerful thermodynamic technique for rapid state preparation and cooling in continuous-variable quantum systems, particularly critical for low-temperature quantum information processing.

quant-ph↗