arXiv subjects

Siyu Zhou

Publications and source records attributed to Siyu Zhou.

At least 19 recordsLinked to original sources

arXiv2026

Neuro-Symbolic Synergy for World Modeling

Large language models (LLMs) exhibit strong general-purpose reasoning capabilities, yet they frequently hallucinate when used as world models (WMs), where strict compliance with deterministic transition rules--particularly in corner cases--is essential. In contrast, Symbolic WMs provide logical consistency but lack semantic expressivity. To bridge this gap, we propose Neuro-Symbolic Synergy (NeSyS), a framework that integrates the probabilistic semantic priors of LLMs with executable symbolic rules to achieve both expressivity and robustness. NeSyS alternates training between the two models using trajectories inadequately explained by the other. Unlike rule-based prompting, the symbolic WM contributes candidate-level scores through log-linear reranking, without requiring the LLM to interpret rule text. Rule-guided sampling prioritizes transitions that are weakly covered by symbolic rules, using 35--60% of the training pairs while outperforming full-data supervised fine-tuning in five of six settings. Experiments on ScienceWorld, WebShop, and PlanCraft demonstrate consistent gains in WM prediction accuracy and data efficiency; one-step lookahead on open-ended WebShop also improves agent reward. Our models, rules, and code are available at https://github.com/tianyi-lab/NeSyS.

cs.CL↗

arXiv2026

Heating Up the Black Hole X-ray Binary Accretion Disk by Superradiance

A superradiant cloud of ultralight axions around a black hole, that is part of an X-ray binary system, can heat up its accretion disk and be detected by the thermal X-ray spectrum emitted by the disk. We consider a derivative coupling of the axions to the plasma fermions and calculate the emissivity of the inverse bremsstrahlung process that results in a temperature fluctuation of the disk. Based on the thin-disk model and the multicolor disk model, we derive the thermal spectrum with axion heating, which shows an enhanced thermal photon flux and a red-/blue- shifted peak spectral frequency. A single bump hunting search of the axion heating signature in the thermal spectrum of a $10M_\odot$ black hole X-ray binary with a spectral measurement sensitivity of 10\% (1\%) can derive the constraint on axion-electron coupling $|g_{ae}|\gtrsim7.5\times 10^{-12} ~(2.4 \times 10^{-12})$ for axion mass $m_a=5.2\times 10^{-12}\,$eV in a saturated $|211\rangle$ state, and $|g_{ae}|\gtrsim4.5\times 10^{-12} ~(1.4 \times 10^{-12})$ for axion mass $m_a=1.0\times10^{-11}\,$eV in a saturated $|322\rangle$ state. The projected sensitivities are competitive with those from XENONnT. A detailed continuum fitting can further improve the detectability and provide a complementary bound to the black hole spin-down measurement.

astro-ph.HE↗

arXiv2026

Deep Research in Physical Sciences: A Multi-Agent Framework and Comprehensive Benchmark

Deep research agents are Large Language Model (LLM)-based systems designed for autonomous, multi-step scientific reasoning, and they hold immense potential for accelerating research in the physical sciences. However, comprehensive and in-depth evaluations of their capabilities within this domain remain lacking. To address this gap, we introduce PhySciBench, a benchmark highly relevant to physical science research, comprising 200 expert-curated questions, balanced between physics and chemistry, across six task categories that reflect real-world scientific workflows. Evaluations of state-of-the-art models and agent systems on PhySciBench reveal limited performance; even the strongest baseline, Gemini Deep Research, achieves an accuracy of only 33.5%. Analysis of failure cases identifies three recurrent deficiencies: fragility in extended reasoning chains, limited knowledge transfer across steps, and a lack of physics-grounded self-verification. Motivated by these findings, we develop DelveAgent, a modular multi-agent framework equipped with an adaptive planning loop, dual-granularity memory, and a hierarchical physics-grounded reflection mechanism. Across four scientific benchmarks, DelveAgent improves accuracy by up to 7.5 percentage points while reducing inference costs to approximately one-third of the strongest baseline. These results establish the significance of PhySciBench as a critical benchmark for evaluating AI systems in the physical sciences and demonstrate that architectural specialization can effectively enhance the reliability of autonomous scientific research. Our data and code are publicly available at https://github.com/yigengjiang/physci-deepresearch.

physics.comp-ph↗

arXiv2026

Point-Wise Geometry-Aware Transformer for Partial-to-Full Point Cloud Registration in Computer-Assisted Surgery

Partial-to-full registration remains challenging due to varying overlap ratios, fluctuating point densities, and the presence of noise. While transformers have shown strong potential for point cloud processing, prior methods typically confine them to global context aggregation, overlooking fine-grained local geometry crucial for accurate correspondence. We propose \emph{GAPR-Net}, a learning-based point cloud registration framework with a coarse-to-fine architecture that combines convolution and transformer modules, in which local and global information is fused between the partial and full point clouds using a cross-attention mechanism. To achieve this, a transformation-invariant point-wise geometric feature representation is proposed, which can robustly capture relative geometric features for individual points with respect to their neighboring points. To evaluate the effectiveness of the proposed approach, experiments are conducted on four geometrically distinct bones, including the tibia, femur, pelvis, and thoracic cartilage. The overall registration recall reaches 94.2\%, the method results in a low RMSE of 1.992 mm and $R^2$ values of 0.908 and 0.974 for rotation and translation, respectively. The results demonstrate that the proposed method effectively addresses the partial-to-full point cloud registration problem. The proposed method enables highly accurate 3D point cloud registration using partial observation, providing a critical foundation for precise surgical navigation and robotic interventions in computer-assisted surgery. The code will be accessed after the double-blind review process.

cs.CV↗

arXiv2026

Adaptive auditing of AI systems with anytime-valid guarantees

A major bottleneck in characterizing the failure modes of generative AI systems is the cost and time of annotation and evaluation. Consequently, adaptive testing paradigms have gained popularity, where one opportunistically decides which cases and how many to annotate based on past results. While this framework is highly practical, its extreme flexibility makes it difficult to draw statistically rigorous conclusions, as it violates classical assumptions: the number of observations is typically limited (often 10 to 50 cases) and decisions regarding sampling and stopping are made in the midst of data collection rather than based a pre-specified rule. To characterize what statistical inferences can be drawn from highly adaptive audits, we introduce a hypothesis testing framework from two 'dueling' perspectives: (i) the model's null that asserts there is no failure mode with performance below a target threshold versus (ii) the auditor's null that asserts they have a sampling strategy that will uncover a failure mode. Leveraging Safe Anytime-Valid Inference (SAVI), we formalize the auditor as conducting 'testing by betting', which translates into simultaneous e-processes for testing the dueling null hypotheses. Furthermore, if the auditor is sufficiently powerful, we prove that these two hypotheses are asymptotically inverses of each other, in that passage of a stringent audit does in fact certify the AI system as being globally robust. Empirically, we demonstrate that our proposed testing procedures maintain anytime-valid type-I error control, outperform pre-specified testing methods, and can reach statistically rigorous conclusions sometimes with as few as 20 observations.

cs.AI↗

arXiv2026

FVG-PT: Adaptive Foreground View-Guided Prompt Tuning for Vision-Language Models

CLIP-based prompt tuning enables pretrained Vision-Language Models (VLMs) to efficiently adapt to downstream tasks. Although existing studies have made significant progress, they pay limited attention to changes in the internal attention representations of VLMs during the tuning process. In this paper, we attribute the failure modes of prompt tuning predictions to shifts in foreground attention of the visual encoder, and propose Foreground View-Guided Prompt Tuning (FVG-PT), an adaptive plug-and-play foreground attention guidance module, to alleviate the shifts. Concretely, FVG-PT introduces a learnable Foreground Reliability Gate to automatically enhance the foreground view quality, applies a Foreground Distillation Compensation module to guide visual attention toward the foreground, and further introduces a Prior Calibration module to mitigate generalization degradation caused by excessive focus on the foreground. Experiments on multiple backbone models and datasets show the effectiveness and compatibility of FVG-PT. Codes are available at: https://github.com/JREion/FVG-PT

cs.CV↗

arXiv2026

Machine learning determines the Mg2SiO4 P-T phase diagram

Phase transitions among Mg2SiO4 and its high-pressure polymorphs (wadsleyite and ringwoodite) are central to mantle dynamics and deep-mantle material cycling. However, the locations and Pressure-Temperature (P-T) dependences of these phase boundaries remain debated, largely due to experimental limitations at extreme conditions and the high computational cost of first-principles free-energy calculations. Here, a machine-learning-potential driven workflow combining non-equilibrium thermodynamic integration (NETI) and two-phase coexistence simulations is employed to enable large-scale, long-timescale molecular dynamics sampling. Within this workflow, the melting curve of forsterite is evaluated and a complete P-T phase diagram is constructed. Relative to conventional ab initio approaches, this strategy reduces computational expense while retaining thermodynamic consistency in phase-stability assessment. The workflow is applicable to efficient evaluation of phase stability and thermodynamic properties in deep-Earth silicate systems.

physics.geo-ph↗

arXiv2025

DreaMontage: Arbitrary Frame-Guided One-Shot Video Generation

The "one-shot" technique represents a distinct and sophisticated aesthetic in filmmaking. However, its practical realization is often hindered by prohibitive costs and complex real-world constraints. Although emerging video generation models offer a virtual alternative, existing approaches typically rely on naive clip concatenation, which frequently fails to maintain visual smoothness and temporal coherence. In this paper, we introduce DreaMontage, a comprehensive framework designed for arbitrary frame-guided generation, capable of synthesizing seamless, expressive, and long-duration one-shot videos from diverse user-provided inputs. To achieve this, we address the challenge through three primary dimensions. (i) We integrate a lightweight intermediate-conditioning mechanism into the DiT architecture. By employing an Adaptive Tuning strategy that effectively leverages base training data, we unlock robust arbitrary-frame control capabilities. (ii) To enhance visual fidelity and cinematic expressiveness, we curate a high-quality dataset and implement a Visual Expression SFT stage. In addressing critical issues such as subject motion rationality and transition smoothness, we apply a Tailored DPO scheme, which significantly improves the success rate and usability of the generated content. (iii) To facilitate the production of extended sequences, we design a Segment-wise Auto-Regressive (SAR) inference strategy that operates in a memory-efficient manner. Extensive experiments demonstrate that our approach achieves visually striking and seamlessly coherent one-shot effects while maintaining computational efficiency, empowering users to transform fragmented visual materials into vivid, cohesive one-shot cinematic experiences.

cs.CV↗

arXiv2025

Axion forces in axion backgrounds

Axions can naturally be very light due to the protection of an (approximate) shift symmetry. Because of their pseudoscalar nature, the long-range force mediated by the axion at tree level is spin dependent, which cannot lead to observable effects between two unpolarized macroscopic objects. At the one-loop level, however, the exchange of two axions does mediate a spin-independent force. This force is coherently enhanced in the presence of an axion background. In this work, we study the two-axion exchange force in a generic axion background. We find that the breaking of the axion shift symmetry plays a crucial role in determining this force. The background-induced axion force $V_{\rm bkg}$ vanishes in the shift-symmetry restoration limit. The shift symmetry can be broken either explicitly by non-perturbative effects or effectively by the axion background. When the shift symmetry is broken, $V_{\rm bkg}$ scales as $1/r$ and could be further enhanced by a large occupation number of the background axions. We investigate possible experimental probes of this effect in two distinct scenarios: an axion dark matter background and a solar axion flux, using fifth-force searches and atomic spectroscopy experiments. In the axion dark matter case, we find that the background-induced axion force can place strong constraints on axion couplings and masses, comparable to existing astrophysical bounds.

hep-ph↗

arXiv2025

Raw Data Matters: Enhancing Prompt Tuning by Internal Augmentation on Vision-Language Models

For CLIP-based prompt tuning, introducing more data as additional knowledge for enhancing fine-tuning process is proved to be an effective approach. Existing data amplification strategies for prompt tuning typically rely on external knowledge (e.g., large language models or pre-structured knowledge bases), resulting in higher costs for data collection and processing, while generally ignoring further utilization of features in image modality. To address this, we propose Augmentation-driven Prompt Tuning (AugPT), a self-contained distillation-based prompt tuning approach using only internal augmentation on raw dataset to better exploit known features. Specifically, AugPT employs self-supervised augmentation on unlabeled images in the training set, and introduces a novel gating mechanism based on consensus test, reusing the pre-trained prompt tuning backbone model to spontaneously filter noisy samples, further enhancing the quality of augmented views. Extensive experiments validate that AugPT simultaneously enhances model performance and generalization capability without using appended external knowledge. The code of AugPT is available at: https://github.com/JREion/AugPT .

cs.CV↗

arXiv2025

I2VControl: Disentangled and Unified Video Motion Synthesis Control

Motion controllability is crucial in video synthesis. However, most previous methods are limited to single control types, and combining them often results in logical conflicts. In this paper, we propose a disentangled and unified framework, namely I2VControl, to overcome the logical conflicts. We rethink camera control, object dragging, and motion brush, reformulating all tasks into a consistent representation based on point trajectories, each managed by a dedicated formulation. Accordingly, we propose a spatial partitioning strategy, where each unit is assigned to a concomitant control category, enabling diverse control types to be dynamically orchestrated within a single synthesis pipeline without conflicts. Furthermore, we design an adapter structure that functions as a plug-in for pre-trained models and is agnostic to specific model architectures. We conduct extensive experiments, achieving excellent performance on various control tasks, and our method further facilitates user-driven creative combinations, enhancing innovation and creativity. Project page: https://wanquanf.github.io/I2VControl .

cs.CV↗

arXiv2025

Phantom-Data : Towards a General Subject-Consistent Video Generation Dataset

Subject-to-video generation has witnessed substantial progress in recent years. However, existing models still face significant challenges in faithfully following textual instructions. This limitation, commonly known as the copy-paste problem, arises from the widely used in-pair training paradigm. This approach inherently entangles subject identity with background and contextual attributes by sampling reference images from the same scene as the target video. To address this issue, we introduce \textbf{Phantom-Data, the first general-purpose cross-pair subject-to-video consistency dataset}, containing approximately one million identity-consistent pairs across diverse categories. Our dataset is constructed via a three-stage pipeline: (1) a general and input-aligned subject detection module, (2) large-scale cross-context subject retrieval from more than 53 million videos and 3 billion images, and (3) prior-guided identity verification to ensure visual consistency under contextual variation. Comprehensive experiments show that training with Phantom-Data significantly improves prompt alignment and visual quality while preserving identity consistency on par with in-pair baselines.

cs.CV↗

arXiv2025

WALL-E 2.0: World Alignment by NeuroSymbolic Learning improves World Model-based LLM Agents

Can we build accurate world models out of large language models (LLMs)? How can world models benefit LLM agents? The gap between the prior knowledge of LLMs and the specified environment's dynamics usually bottlenecks LLMs' performance as world models. To bridge the gap, we propose a training-free "world alignment" that learns an environment's symbolic knowledge complementary to LLMs. The symbolic knowledge covers action rules, knowledge graphs, and scene graphs, which are extracted by LLMs from exploration trajectories and encoded into executable codes to regulate LLM agents' policies. We further propose an RL-free, model-based agent "WALL-E 2.0" through the model-predictive control (MPC) framework. Unlike classical MPC requiring costly optimization on the fly, we adopt an LLM agent as an efficient look-ahead optimizer of future steps' actions by interacting with the neurosymbolic world model. While the LLM agent's strong heuristics make it an efficient planner in MPC, the quality of its planned actions is also secured by the accurate predictions of the aligned world model. They together considerably improve learning efficiency in a new environment. On open-world challenges in Mars (Minecraft like) and ALFWorld (embodied indoor environments), WALL-E 2.0 significantly outperforms existing methods, e.g., surpassing baselines in Mars by 16.1%-51.6% of success rate and by at least 61.7% in score. In ALFWorld, it achieves a new record 98% success rate after only 4 iterations.

cs.AI↗

arXiv2025

Phantom: Subject-consistent video generation via cross-modal alignment

The continuous development of foundational models for video generation is evolving into various applications, with subject-consistent video generation still in the exploratory stage. We refer to this as Subject-to-Video, which extracts subject elements from reference images and generates subject-consistent videos following textual instructions. We believe that the essence of subject-to-video lies in balancing the dual-modal prompts of text and image, thereby deeply and simultaneously aligning both text and visual content. To this end, we propose Phantom, a unified video generation framework for both single- and multi-subject references. Building on existing text-to-video and image-to-video architectures, we redesign the joint text-image injection model and drive it to learn cross-modal alignment via text-image-video triplet data. The proposed method achieves high-fidelity subject-consistent video generation while addressing issues of image content leakage and multi-subject confusion. Evaluation results indicate that our method outperforms other state-of-the-art closed-source commercial solutions. In particular, we emphasize subject consistency in human generation, covering existing ID-preserving video generation while offering enhanced advantages.

cs.CV↗

arXiv2025

MAO: Efficient Model-Agnostic Optimization of Prompt Tuning for Vision-Language Models

Though CLIP-based prompt tuning significantly enhances pre-trained Vision-Language Models, existing research focuses on reconstructing the model architecture, e.g., additional loss calculation and meta-networks. These approaches generally lead to increased complexity and extended training cost. To maintain the efficiency of the tuning process, we propose plug-and-play Model-Agnostic Optimization (MAO) for prompt tuning. Without altering any components of the prompt tuning backbone, we introduce a Data-Driven Enhancement framework to optimize the distribution of the initial data, and incorporate an Alterable Regularization module to boost the task-specific feature processing pipeline, thereby improving overall performance while maintaining low computational cost. Extensive experiments on MAO demonstrate its outstanding performance and efficiency. The code of MAO is available at: https://github.com/JREion/M.A.O .

cs.CV↗

arXiv2025

I2VControl-Camera: Precise Video Camera Control with Adjustable Motion Strength

Video generation technologies are developing rapidly and have broad potential applications. Among these technologies, camera control is crucial for generating professional-quality videos that accurately meet user expectations. However, existing camera control methods still suffer from several limitations, including control precision and the neglect of the control for subject motion dynamics. In this work, we propose I2VControl-Camera, a novel camera control method that significantly enhances controllability while providing adjustability over the strength of subject motion. To improve control precision, we employ point trajectory in the camera coordinate system instead of only extrinsic matrix information as our control signal. To accurately control and adjust the strength of subject motion, we explicitly model the higher-order components of the video trajectory expansion, not merely the linear terms, and design an operator that effectively represents the motion strength. We use an adapter architecture that is independent of the base model structure. Experiments on static and dynamic scenes show that our framework outperformances previous methods both quantitatively and qualitatively. The project page is: https://wanquanf.github.io/I2VControlCamera .

cs.CV↗

arXiv2024

Precise and Fast LIDAR via Electrical Asynchronous Sampling Based on a Single Femtosecond Laser

Using a laser-based ranging method for precise environmental 3D sensing, LiDAR has numerous applications in science and industry. However, conventional LiDAR face challenges in simultaneously achieving high ranging precision and fast measurement rates, which limits their applicability in more precise fields, such as aerospace, smart healthcare and beyond. By employing an asynchronous electrical pulse sampling strategy on a single optical frequency comb with a stable repetition rate and femtosecond-pulse width, we exploit the advantages of optical-frequency-comb ranging method and overcome the limitations of sampling aliasing and low data-utilization inherent in traditional approaches. This enables a significant improvement of LiDAR's performance to achieve micrometer-level precision and megahertz-regimes update rates over meter-range on non-cooperative targets. Specifically, we achieve 38.8-$μ$m Allan deviation at 1-MHz update rate and 8.06-$μ$m Allan deviation after 2-ms time-averaging based on a 56.091-MHz femtosecond laser. This enhancement enables various advanced measurement applications, including metrology monitoring on high-speed objects, 1-megapixel/s precise 3D scanning imaging and first-ever contactless vital sign detection using time-of-flight LiDAR. This LiDAR unlock new possibilities for precise and fast real-time measurements in diverse fields.

physics.optics↗

arXiv2024

Atomic evolution of hydrogen intercalation wave dynamics in palladium nanocrystals

Solute-intercalation-induced phase separation creates spatial heterogeneities in host materials, a phenomenon ubiquitous in batteries, hydrogen storage, and other energy devices. Despite many efforts, probing intercalation processes at the atomic scale has been a significant challenge. We study hydrogen (de)intercalation in palladium nanocrystals as a model system and achieve atomic-resolution imaging of hydrogen intercalation wave dynamics by utilizing liquid-phase transmission electron microscopy. Our observations reveal that intercalation wave mechanisms, instead of shrinking-core mechanisms, prevail at ambient temperature for palladium nanocubes ranging from ~60 nm down to ~10 nm. We uncover the atomic evolution of hydrogen intercalation wave transitioning from non-planar and inclined boundaries to those closely aligned with {100} planes. Our kinetic Monte Carlo simulations demonstrate the observed intercalation wave dynamics correspond to sorption pathways minimizing the lattice mismatch strain at the phase boundary. Unveiling the atomic intercalation pathways holds profound implications for engineering intercalation-mediated devices and advancements in energy sciences.

cond-mat.stat-mech↗