arXiv subjects

Yunlong Liu

Publications and source records attributed to Yunlong Liu.

At least 19 recordsLinked to original sources

arXiv2026

Finite-temperature transition-window weights on a fault-tolerant quantum computer: duplex filtering on a thermal purification

The excite-filter-readout framework of Kharazi et al. computes zero-temperature response windows on a fault-tolerant quantum computer from a pure ground state. We extend it to Gibbs states, where the two objects decouple: filters act on absolute energies, a thermal response on transition frequencies. The extension rests on a Liouville-space/TFD identity, explicit in the block-encoding/QSP setting: the thermally weighted transition-window weight is a spectral projection of H_0 (x) I - I (x) H_0^T onto the Schmidt (thermofield-double) purification, instance of Rall's framework. Around it we assemble a routed-SELECT block encoding (one Hamiltonian query per filter query), a conjugate-pair readout, and a four-route (duplex, bin, Fourier-LCU, METTS) compatibility analysis. The cost is multiplicative and conditional: a per-window query bound M_"AE" * (2*sqrt(d/Z_"lb")*d_"vt" + 4*d_W), with a tight signal-adaptive count, the optimal sqrt(alpha beta) degree, and a window-edge term. The preparation factor sqrt(d/Z) grows exponentially with system size at fixed temperature in positive-entropy-density systems (2.08 -> 25.7, N = 8 to 14 at beta = 2), relieved only as beta -> 0. Zero temperature is recovered for beta G >= ln(2(d-1)||D||^2/epsilon). Identities are cross-checked between independent formula paths to 10^(-16)-10^(-12) and the readout assembly is validated at matrix level under a block-encoding promise.

quant-ph↗

arXiv2026

Quantum Algorithms for Molecular Response Functions: Resolvent Kernels, Time-Domain Sampling, Query Lower Bounds with Matched Single-Point Calibers, and a Six-Route Comparison

Quantum algorithms for molecular response functions have mostly been studied one route at a time, most prominently the indicator-binning algorithm of Kharazi et al. We compare six routes within one accounting framework and state the conversions between their resource calibers. On the resolvent route the Lorentz-broadened spectrum is exactly a resolvent expectation; at finite gamma two boundary terms, a reversal term and a permanent-dipole pseudo-line, enter a disjoint error budget. Because the kernel subnormalization is exactly 1, the published binning accounting carries a removable Theta(alpha+1+omega) factor (49.69x and 7.75x at the definitional caliber), a correction that applies to both routes. Estimating the square root of the kernel amplitude lowers the readout cost from beta^2/mu_tilde_2 to beta/sqrt(mu_tilde_2). On the time-domain route we prove a single-point amplitude-estimation theorem and convert the generalized QPE of Loaiza et al. to the same accounting. The lower bounds are Theta_tilde(1/gamma) for parameter location, Theta(1/(gamma epsilon)) for windowed values, Omega(n/gamma) inside a restricted correlation-sampling class, and Omega(1/(epsilon gamma^2)) on the Lorentzian object; in the single-circuit-resolution caliber, gamma^(-n) is a grid-coverage artifact. Six routes, their calibers, an n=2 assembly with a component-grade coherence, and a phase diagram over the tested ensembles form one accounting map.

quant-ph↗

arXiv2026

SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning

Agentic reinforcement learning (RL) has emerged as an important post-training approach for enhancing the capabilities of Large Language Models (LLMs). However, existing methods face a trade-off between policy performance and resource efficiency. Conventional Proximal Policy Optimization (PPO) implementations incur substantial memory overhead from a separate critic, whereas critic-free group-relative methods require multiple rollouts and face potential learning bottlenecks on long-horizon tasks. In this work, we propose Single-rollout Autoregressive Policy Optimization (SAPO), an efficient PPO-style framework that unifies policy optimization and value learning within a single causal language model. SAPO exploits the autoregressive structure of LLMs to sequentially generate action and value estimation at distinct causal boundaries with shared parameters, and then jointly optimizes the PPO objectives and an auxiliary on-policy SARSA objective with turn-level generalized advantage estimation, where the latter is designed to facilitate value learning. Extensive experiments on ALFWorld and WebShop with Qwen2.5-1.5B/7B and Qwen3-14B demonstrate that SAPO reduces peak GPU memory usage by 24.8% and per-iteration runtime by 23.1% over strong PPO baseline, while matching or slightly improving task success rate. Our experiments also show that SAPO outperforms Group Relative Policy Optimization (GRPO) and recent cutting-edge variants in both task success and training stability.

cs.AI↗

arXiv2026

Vanishing Altermagnetism and Emergent Ferromagnetism at the Two-Dimensional Limit of CrSb

Altermagnets offer a promising route to spintronic devices by combining spin-split electronic states with zero net magnetization, but how reduced dimensionality reshapes their magnetic order remains poorly understood. Here, we grow bilayer-thickness CrSb (0001) films directly on SrTiO3(001) substrates by molecular beam epitaxy and uncover an unexpected thickness-dependent magnetic transition. The one-bilayer films exhibit room-temperature ferromagnetism with strong perpendicular magnetic anisotropy, whereas ferromagnetism vanishes in two-bilayer films, contrasting sharply with the altermagnetic ground state of bulk CrSb. Density functional theory calculations reveal that the exposed Cr surface induces orbital rearrangement that strengthens intralayer ferromagnetic exchange while suppressing competing interlayer antiferromagnetic coupling, driving the dimensional crossover from altermagnetism to ferromagnetism. These findings establish dimensionality and surface termination as powerful controls of magnetic order and provide a platform for engineering ultrathin magnetic states for spintronic applications.

cond-mat.mtrl-sci↗

arXiv2026

PlanPO: Group Planning-Aware Policy Optimization for Multi-Turn Agentic LLMs

Group-relative policy optimization has emerged as a key paradigm for training agentic large language models (LLMs) on multi-turn interactive tasks. However, most existing variants fail to distinguish advantages among successful trajectories even when these trajectories differ substantially in their interaction efficiency. For instance, circuitous successes are often assigned the identical outcome reward, causing advantage collapse and severe performance bottlenecks. To this end, we propose Group Planning-aware Policy Optimization (PlanPO), a simple yet effective RL method for learning generalizable planning abilities beyond task-specific high-quality behavior patterns. Specifically, PlanPO introduces coarse-to-fine advantage signals, which capture the relative differences in trajectory-level lengths and turn-level response lengths conditioned on successful trajectories sampled for the same task. Within the group-relative optimization structure, this enables agents to actively learn generalizable and deliberate behaviors spanning interaction planning and textual generation from high-quality rollouts, without degenerating into vanilla length minimization. Experimentally, PlanPO improves over GRPO by 27.2\% on average across the challenging multi-turn benchmarks ALFWorld, WebShop, and SciWorld, outperforming recent powerful baselines while incurring negligible additional training cost.

cs.AI↗

arXiv2026

Content-Induced Spatial-Spectral Aggregation Network for Change Detection in Remote Sensing Images

The integration of spatial and spectral information is beneficial to the improvement of change detection performance. However, existing methods cannot efficiently suppress the influences of spatial and spectral differences in unchanged areas. To address these issues, in this paper we propose a content-guided spatial-spectral integration network (CSI-Net) for the fusion of global spatial details and spectral difference information. Specifically, the proposed CSI-Net is composed of a spatial reasoning (SR) module, a spectral difference (SD) module, and a content-guided integration (CGI) module. In the SR module, the spatial information is learned by cascaded graph convolution blocks for global modeling. The SD module is responsible for the extraction of spectral features, by calculating the means and variances of features to reduce the impact of spectral differences in unchanged regions. In addition, in order to integrate the spatial-spectral features efficiently, we design a CGI module to further take advantage of their complementary information. In this module, high-level content information is introduced as a guide for a proper interaction. Due to the efficient spatial-spectral fusion, the proposed CSI-Net can learn the changed features better while achieving a suppression of spectral differences. Experimental results on LEVIR-CD, WHU-CD, and CLCD datasets demonstrate that the proposed CSI-Net produces better performance compared to state-of-the-art methods, and is applicable to different scenarios

cs.CV↗

arXiv2026

Building Change Detection in Earthquake: A Multi-Scale Interaction Network and A Change Detection Dataset

As one of the most destructive natural disasters, earthquakes have struck many countries around the world in recent years, causing serious economic losses. Change detection (CD) can be applied to post-earthquake damage assessment as it can infer destroyed change regions from multi-temporal remote sensing images. Furthermore, the CD with short imaging interval will better satisfy the needs of the emergency rescues after earthquakes. However, the capability of current methods built on deep neural networks is limited because the dataset with short imaging interval is absent. To meet post-disaster immediate relief, we create a CD dataset, Turkey earthquake CD dataset (TUE-CD), for the evaluation of building damage in the short term after an earthquake. Because of the short acquisition interval of the post-event images, the imaging angle is different for different temporal images, which leads to some side-looking problems. To deal with these challenges, we present a multi-scale feature interaction network (MSI-Net) for efficient interaction between bi-temporal features, as well as mitigating the effect of side-looking problems. Specifically, the proposed MSI-Net consists of joint cross-attention (JCA) modules, multi-scale offset calibration (MOC) modules, and feature integration (FeI) modules. The JCA module unifies channel cross-attention and spatial joint attention for sufficient feature interaction. The MOC module further estimates the offsets to align the bi-temporal image with the multi-scale features. Finally, calibrated features and multi-scale features are fused by FeI modules for the prediction of changed areas. Experiments on the WHU-CD, CLCD, and the constructed TUE-CD dataset indicate that the proposed MSI-Net provides better results than considered state-of-the-art CD methods.

cs.CV↗

arXiv2026

FMRFusion: Frequency-Aware Multi-View Representation Learning for Heterogeneous Image Fusion

Infrared and visible image fusion aims to generate a composite image that retains significant target information and preserves detailed textures, integrating two heterogeneous modalities. Previous image fusion methods typically adopt a single-module stacking approach to extract features from the two modalities. However, these approaches may result in incomplete learning of their distinct characteristics, thereby limiting the fusion effectiveness and constrain ing robustness in real-world heterogeneous data scenarios. To address these challenges, we propose FMRFusion, a frequency-aware multi-view representation learning network for Heterogeneous Image Fusion. A Multi-Scale Struc tural Perception Module is introduced to effectively capture discriminative structures, extracting fine-grained local structures and essential contextual information. A bilinear frequency decomposition mechanism is employed to sepa rate features into high-frequency and low-frequency components, enabling joint modeling of local details and global representations across different frequency domains. Moreover, a Cross-View Complementary Interaction is incorpo rated to explicitly model and fuse the complementary characteristics between reflected light information and radiative intensity responses, facilitating effective cross-view interaction. We further improve the Performance of the fused results by flow matching, which progressively refines the fused features by learning the transformation from coarse data to high-quality representations. Extensive experiments conducted on multiple benchmark datasets demonstrate that FMRFusion achieves superior and consistent performance across a range of fusion tasks, especially in nighttime scenarios

cs.CV↗

arXiv2026

Periodic orbits and gravitational waveforms of black holes in bumblebee gravity

In this paper, we investigate the dynamics of massive particles and the associated gravitational waveforms in the spacetime of a black hole within the framework of Einstein-Bumblebee gravity. Our analysis encompasses both charged and uncharged black hole configurations, with a particular focus on the spontaneous Lorentz symmetry breaking mechanism inherent to this model, which is governed by a dimensionless coupling parameter $l$. We analyze the geodesic equations and the effective potential to determine the allowed parameter space for bound orbits, demonstrating that in the charged case, both the Lorentz-violating parameter $l$ and the electric charge $Q$ significantly enhance the confinement capacity of the potential, thereby broadening the energy and angular momentum windows for bound states. A key focus is placed on the classification and properties of periodic orbits, characterized by rational frequency ratios using the whirl, zoom, and vertex taxonomy. We demonstrate that in the uncharged case ($Q=0$), the radial effective potential and standard innermost stable circular orbit (ISCO) properties are degenerate with those of a Schwarzschild black hole. However, despite this degeneracy in static potential properties, the structure of periodic orbits exhibits qualitative differences, providing a possible observational signature that can break this degeneracy. Finally, we compute the corresponding gravitational waveforms extracted from these periodic orbits using the quadrupole formula. The results reveal that $l$ and $Q$ introduce contrasting phase-shifting effects on the waveforms. This suggests that bumblebee gravity leaves measurable imprints on gravitational-wave signals that could be detected by future space-based gravitational-wave observatories.

gr-qc↗

arXiv2026

InterReal: A Unified Physics-Based Imitation Framework for Learning Human-Object Interaction Skills

Interaction is one of the core abilities of humanoid robots. However, most existing frameworks focus on non-interactive whole-body control, which limits their practical applicability. In this work, we develop InterReal, a unified physics-based imitation learning framework for Real-world human-object Interaction (HOI) control. InterReal enables humanoid robots to track HOI reference motions, facilitating the learning of fine-grained interactive skills and their deployment in real-world settings. Within this framework, we first introduce a HOI motion data augmentation scheme with hand-object contact constraints, and utilize the augmented motions to improve policy stability under object perturbations. Second, we propose an automatic reward learner to address the challenge of large-scale reward shaping. A meta-policy guided by critical tracking error metrics explores and allocates reward signals to the low-level reinforcement learning objective, which enables more effective learning of interactive policies. Experiments on HOI tasks of box-picking and box-pushing demonstrate that InterReal achieves the best tracking accuracy and the highest task success rate compared to recent baselines. Furthermore, we validate the framework on the real-world robot Unitree G1, which demonstrates its practical effectiveness and robustness beyond simulation.

cs.RO↗

arXiv2026

Task-Aware Exploration via a Predictive Bisimulation Metric

Accelerating exploration in visual reinforcement learning under sparse rewards remains challenging due to the substantial task-irrelevant variations. Despite advances in intrinsic exploration, many methods either assume access to low-dimensional states or lack task-aware exploration strategies, thereby rendering them fragile in visual domains. To bridge this gap, we present TEB, a Task-aware Exploration approach that tightly couples task-relevant representations with exploration through a predictive Bisimulation metric. Specifically, TEB leverages the metric not only to learn behaviorally grounded task representations but also to measure behaviorally intrinsic novelty over the learned latent space. To realize this, we first theoretically mitigate the representation collapse of degenerate bisimulation metrics under sparse rewards by internally introducing a simple but effective predicted reward differential. Building on this robust metric, we design potential-based exploration bonuses, which measure the relative novelty of adjacent observations over the latent space. Extensive experiments on MetaWorld and Maze2D show that TEB achieves superior exploration ability and outperforms recent baselines.

cs.AI↗

arXiv2026

Periodic orbits and gravitational waveforms of spinning particles in nonlocal Gravity

In this paper, we investigate the dynamics and gravitational-wave signatures of periodic orbits of spinning test particles moving in the equatorial plane around static, spherically symmetric black holes within the framework of Deser-Woodard nonlocal gravity. Based on the Mathisson-Papapetrou-Dixon equations, combined with the Tulczyjew spin supplementary condition, we derive the orbital dynamic equations for spinning particles moving in the equatorial plane and impose a timelike constraint to exclude unphysical superluminal trajectories. By comparing with the classical Schwarzschild black hole, we systematically analyze the effects of the nonlocal gravitational parameters $ζ$ and $b$ on the effective potential governing the radial motion of particles and the innermost stable circular orbit. In addition, gravitational waveforms exhibit significant phase differences: an increase in $ζ$ induces a phase delay, whereas an increase in $b$ results in a phase advance. A one-year simulation of the orbital evolution of an extreme mass ratio inspiral demonstrates that when $b=2$ and $ζ\approx10^{-6}$, the mismatch between the gravitational waveforms predicted for the nonlocal gravity black hole and those for the Schwarzschild black hole reaches the distinguishable threshold ($\mathcal{M}=0.0125$), providing a basis for observational discrimination between general relativity and nonlocal gravity.

gr-qc↗

arXiv2025

From Pixels to Predicates Structuring urban perception with scene graphs

Perception research is increasingly modelled using streetscapes, yet many approaches still rely on pixel features or object co-occurrence statistics, overlooking the explicit relations that shape human perception. This study proposes a three stage pipeline that transforms street view imagery (SVI) into structured representations for predicting six perceptual indicators. In the first stage, each image is parsed using an open-set Panoptic Scene Graph model (OpenPSG) to extract object predicate object triplets. In the second stage, compact scene-level embeddings are learned through a heterogeneous graph autoencoder (GraphMAE). In the third stage, a neural network predicts perception scores from these embeddings. We evaluate the proposed approach against image-only baselines in terms of accuracy, precision, and cross-city generalization. Results indicate that (i) our approach improves perception prediction accuracy by an average of 26% over baseline models, and (ii) maintains strong generalization performance in cross-city prediction tasks. Additionally, the structured representation clarifies which relational patterns contribute to lower perception scores in urban scenes, such as graffiti on wall and car parked on sidewalk. Overall, this study demonstrates that graph-based structure provides expressive, generalizable, and interpretable signals for modelling urban perception, advancing human-centric and context-aware urban analytics.

cs.CV↗

arXiv2025

Integration of nested cross-validation, automated hyperparameter optimization, high-performance computing to reduce and quantify the variance of test performance estimation of deep learning models

Background and Objectives: The variability and biases in the real-world performance benchmarking of deep learning models for medical imaging compromise their trustworthiness for real-world deployment. The common approach of holding out a single fixed test set fails to quantify the variance in the estimation of test performance metrics. This study introduces NACHOS (Nested and Automated Cross-validation and Hyperparameter Optimization using Supercomputing) to reduce and quantify the variance of test performance metrics of deep learning models. Methods: NACHOS integrates Nested Cross-Validation (NCV) and Automated Hyperparameter Optimization (AHPO) within a parallelized high-performance computing (HPC) framework. NACHOS was demonstrated on a chest X-ray repository and an Optical Coherence Tomography (OCT) dataset under multiple data partitioning schemes. Beyond performance estimation, DACHOS (Deployment with Automated Cross-validation and Hyperparameter Optimization using Supercomputing) is introduced to leverage AHPO and cross-validation to build the final model on the full dataset, improving expected deployment performance. Results: The findings underscore the importance of NCV in quantifying and reducing estimation variance, AHPO in optimizing hyperparameters consistently across test folds, and HPC in ensuring computational feasibility. Conclusions: By integrating these methodologies, NACHOS and DACHOS provide a scalable, reproducible, and trustworthy framework for DL model evaluation and deployment in medical imaging. To maximize public availability, the full open-source codebase is provided at https://github.com/thepanlab/NACHOS

cs.CV↗

arXiv2025

Solar-system experimental constraints on nonlocal gravity

In this work, we study the constraints on the characteristic parameters $(ζ,b)$ of the Deser-Woodard nonlocal gravity model in a static and spherically symmetric background, using four classes of high-precision Solar-System experiments: stellar light deflection, Shapiro time delay, perihelion advance, and geodetic precession. From geodesic equations, we derive observable geometric quantities that can be directly compared with VLBI/VLBA astrometry, the Cassini time-delay measurement, MESSENGER data and the GP-B/LLR results. Our results show that a larger value of $b$ suppresses the nonlocal effect more rapidly with radius, thereby weakening the overall constraints on $ζ$. The perihelion advance exhibits the strongest sensitivity to $ζ$ around $b\simeq 1.06$, providing the tightest single experiment bound, whereas away from this region the combined constraint becomes dominated by the Shapiro time delay. Incorporating all four experiments yields a well-defined and sharply bounded allowed region for the parameter space $(ζ,b)$.

gr-qc↗

arXiv2025

Quasinormal modes and tidal Love numbers of covariant effective quantum black holes with cosmological constant

This paper systematically investigates the quasinormal modes (QNMs) and tidal Love numbers of covariant quantum-modified black holes (BHs) within two cosmological constant-dependent metric frameworks. By deriving axial/polar perturbation equations and employing pseudospectral methods (for QNMs) and numerical integration (for Love numbers), we quantify the quantum parameter $ζ$'s influence. We find that, higher overtones exhibit enhanced $ζ$ sensitivity. Solution 1 maintains purely imaginary QNMs under $ζ$ variation, whereas Solution 2 demonstrates mode transitions from real to complex for polar/axial perturbations. Furthermore, for Love number, axial perturbations show non-monotonic $ζ$ dependence, with $Δ\mathcal{T}^a_l$ peaking at specific $\ell$ values before decay. While for polar perturbation of the same metric displays strictly monotonic $ζ$ effects on $Δ\mathcal{T}^p_l$. These results highlight quantum gravity's role in BH perturbation dynamics and cosmological BH phenomenology.

gr-qc↗

arXiv2025

Tunable spin-wave nonreciprocity in ferrimagnetic domain-wall channels

The nonreciprocal propagation of spin waves (SWs) offers opportunities for developing novel functional magnonic logic devices, where controllability is crucial for magnetic signal processing. Domain walls act as natural waveguides due to their magnetic configuration, offering a platform for the in-depth investigation of nonreciprocal SW propagation and its manipulation. In this work, we theoretically and numerically investigate the tunable spin-wave nonreciprocity in ferrimagnetic domain-wall channels under the influence of an external field. It is revealed that the Dzyaloshinskii-Moriya interaction exerts dual control over both nonreciprocal spin-wave propagation and spin-splitting phenomena. Moreover, SW nonreciprocity is magnetically tunable, with its sign reversibly switched by inverting the applied field direction, while preserving the host spin configuration. The orientation of the magnetic field can selectively stabilize or destabilize the domain wall structure, offering precise control over spin-wave nonreciprocity. Ultimately, we demonstrate a controllable SW transmission scheme via external magnetic field modulation, providing critical insights for the design of future magnonic devices.

cond-mat.mtrl-sci↗

arXiv2025

CrossLinear: Plug-and-Play Cross-Correlation Embedding for Time Series Forecasting with Exogenous Variables

Time series forecasting with exogenous variables is a critical emerging paradigm that presents unique challenges in modeling dependencies between variables. Traditional models often struggle to differentiate between endogenous and exogenous variables, leading to inefficiencies and overfitting. In this paper, we introduce CrossLinear, a novel Linear-based forecasting model that addresses these challenges by incorporating a plug-and-play cross-correlation embedding module. This lightweight module captures the dependencies between variables with minimal computational cost and seamlessly integrates into existing neural networks. Specifically, it captures time-invariant and direct variable dependencies while disregarding time-varying or indirect dependencies, thereby mitigating the risk of overfitting in dependency modeling and contributing to consistent performance improvements. Furthermore, CrossLinear employs patch-wise processing and a global linear head to effectively capture both short-term and long-term temporal dependencies, further improving its forecasting precision. Extensive experiments on 12 real-world datasets demonstrate that CrossLinear achieves superior performance in both short-term and long-term forecasting tasks. The ablation study underscores the effectiveness of the cross-correlation embedding module. Additionally, the generalizability of this module makes it a valuable plug-in for various forecasting tasks across different domains. Codes are available at https://github.com/mumiao2000/CrossLinear.

cs.LG↗