arXiv subjects

Wei Li

Publications and source records attributed to Wei Li.

At least 343 records · Page 19Linked to original sources

arXiv2025

Enhancing Physical Consistency in Lightweight World Models

A major challenge in deploying world models is the trade-off between size and performance. Large world models can capture rich physical dynamics but require massive computing resources, making them impractical for edge devices. Small world models are easier to deploy but often struggle to learn accurate physics, leading to poor predictions. We propose the Physics-Informed BEV World Model (PIWM), a compact model designed to efficiently capture physical interactions in bird's-eye-view (BEV) representations. PIWM uses Soft Mask during training to improve dynamic object modeling and future prediction. We also introduce a simple yet effective technique, Warm Start, for inference to enhance prediction quality with a zero-shot model. Experiments show that at the same parameter scale (400M), PIWM surpasses the baseline by 60.6% in weighted overall score. Moreover, even when compared with the largest baseline model (400M), the smallest PIWM (130M Soft Mask) achieves a 7.4% higher weighted overall score with a 28% faster inference speed.

cs.AI↗

arXiv2025

Antiferromagnetic ordering and critical behavior induced giant magnetocaloric effect in distorted kagome lattice Gd$_3$BWO$_9$

We synthesize the high-quality Gd$_3$BWO$_9$ single crystal and investigate its lowtemperature magnetic and thermodynamic properties. Below $T\rm_{N}$ = 1.08 K, the anisotropic behavior of magnetic susceptibilities reveals that the Gd$^{3+}$ moments exhibit the dominant antiferromagnetic coupling along the $c$-axis, while displaying a ferromagnetic arrangement in kagome plane. With pronounced magnetic frustration, in adiabatic demagnetization refrigeration experiments starting from initial conditions of 9 T and 2 K, Gd$_3$BWO$_9$ polycrystal reaches a minimum temperature of 0.151 K, significantly lower than its $T\rm_{N}$. Due to the high density of Gd$^{3+}$ ions ($S$=7/2), the maximum magnetic entropy change reaches over 50 J kg$^{-1}$ K$^{-1}$ under fields up to 7 T in Gd$_3$BWO$_9$, nearly 1.5 times as large as commercial sub-Kelvin magnetic coolant Gd$_3$Ga$_5$O$_{12}$(GGG). The H-T phase diagram of Gd$_3$BWO$_9$ under $H$//$c$ exhibits field-induced critical behavior near the phase boundaries. This observation aligns with the theoretical scenario in which a quantum critical point acts as the endpoint of a line of classical second-order phase transitions. Such behavior suggests the importance of further investigations into the divergence of magnetic Grüneisen parameter in the vicinity of critical field at ultralow temperatures.

cond-mat.str-el↗

arXiv2025

Supervised and unsupervised learning with numerical computation for the Wolfram cellular automata

The local rules of Wolfram cellular automata with one-dimensional three-cell neighborhoods are represented by eight-bit binary that encode deterministic update rules. These automata are widely utilized to investigate self-organization phenomena and the dynamics of complex systems. In this work, we employ numerical simulations and computational methods to investigate the asymptotic density and dynamical evolution mechanisms in Wolfram automata. We apply both supervised and unsupervised learning methods to identify the configurations associated with different Wolfram rules. Furthermore, we explore alternative initial conditions under which certain Wolfram rules generate similar fractal patterns over time, even when starting from a single active site. Our results reveal the relationship between the asymptotic density and the initial density of selected rules. The supervised learning methods effectively identify the configurations of various Wolfram rules, while unsupervised methods like principal component analysis and autoencoders can approximately cluster configurations of different Wolfram rules into distinct groups, yielding results that align well with simulated density outputs.

physics.comp-ph↗

arXiv2025

Robix: A Unified Model for Robot Interaction, Reasoning and Planning

We introduce Robix, a unified model that integrates robot reasoning, task planning, and natural language interaction within a single vision-language architecture. Acting as the high-level cognitive layer in a hierarchical robot system, Robix dynamically generates atomic commands for the low-level controller and verbal responses for human interaction, enabling robots to follow complex instructions, plan long-horizon tasks, and interact naturally with human within an end-to-end framework. Robix further introduces novel capabilities such as proactive dialogue, real-time interruption handling, and context-aware commonsense reasoning during task execution. At its core, Robix leverages chain-of-thought reasoning and adopts a three-stage training strategy: (1) continued pretraining to enhance foundational embodied reasoning abilities including 3D spatial understanding, visual grounding, and task-centric reasoning; (2) supervised finetuning to model human-robot interaction and task planning as a unified reasoning-action sequence; and (3) reinforcement learning to improve reasoning-action consistency and long-horizon task coherence. Extensive experiments demonstrate that Robix outperforms both open-source and commercial baselines (e.g., GPT-4o and Gemini 2.5 Pro) in interactive task execution, demonstrating strong generalization across diverse instruction types (e.g., open-ended, multi-stage, constrained, invalid, and interrupted) and various user-involved tasks such as table bussing, grocery shopping, and dietary filtering.

cs.AI↗

arXiv2025

Kitaev-derived Gapless Spin Liquid in the $K$-$J$-$Γ$-$Γ'$ Quantum Magnet Na$_2$Co$_2$TeO$_6$

The realization of quantum spin liquids (QSLs) in Kitaev magnets represents an intriguing topic in frustrated quantum magnetism. Despite prediction in the pure Kitaev honeycomb model, realization of QSLs in realistic systems and materials remain scarce. The recent discovery of cobalt-based compound Na$_2$Co$_2$TeO$_6$ has raised significant research interest. By establishing a realistic $K$-$J$-$Γ$-$Γ'$ model for Na$_2$Co$_2$TeO$_6$ -- with a dominant antiferromagnetic (AFM) Kitaev interaction ($K>0$) that quantitatively explains its thermodynamics measurements -- we reveal an intermediate gapless QSL phase under [111] magnetic fields with tensor-network calculations. We confirm the QSL nature of this phase by demonstrating its adiabatic connection to the intensively studied intermediate QSL of the pure AFM Kitaev model under out-of-plane fields. Our results show excellent agreement with recent high-field experiments, thereby explaining the intermediate-field phase in Na$_2$Co$_2$TeO$_6$. These findings bridge the gap between theoretical proposals for a Kitaev-derived QSL and experimental realization, opening new avenues for exploring exotic quantum states of matter in realistic Kitaev materials.

cond-mat.str-el↗

arXiv2025

HingeNet: A Harmonic-Aware Fine-Tuning Approach for Beat Tracking

Fine-tuning pre-trained foundation models has made significant progress in music information retrieval. However, applying these models to beat tracking tasks remains unexplored as the limited annotated data renders conventional fine-tuning methods ineffective. To address this challenge, we propose HingeNet, a novel and general parameter-efficient fine-tuning method specifically designed for beat tracking tasks. HingeNet is a lightweight and separable network, visually resembling a hinge, designed to tightly interface with pre-trained foundation models by using their intermediate feature representations as input. This unique architecture grants HingeNet broad generalizability, enabling effective integration with various pre-trained foundation models. Furthermore, considering the significance of harmonics in beat tracking, we introduce harmonic-aware mechanism during the fine-tuning process to better capture and emphasize the harmonic structures in musical signals. Experiments on benchmark datasets demonstrate that HingeNet achieves state-of-the-art performance in beat and downbeat tracking

cs.SD↗

arXiv2025

BeatFM: Improving Beat Tracking with Pre-trained Music Foundation Model

Beat tracking is a widely researched topic in music information retrieval. However, current beat tracking methods face challenges due to the scarcity of labeled data, which limits their ability to generalize across diverse musical styles and accurately capture complex rhythmic structures. To overcome these challenges, we propose a novel beat tracking paradigm BeatFM, which introduces a pre-trained music foundation model and leverages its rich semantic knowledge to improve beat tracking performance. Pre-training on diverse music datasets endows music foundation models with a robust understanding of music, thereby effectively addressing these challenges. To further adapt it for beat tracking, we design a plug-and-play multi-dimensional semantic aggregation module, which is composed of three parallel sub-modules, each focusing on semantic aggregation in the temporal, frequency, and channel domains, respectively. Extensive experiments demonstrate that our method achieves state-of-the-art performance in beat and downbeat tracking across multiple benchmark datasets.

cs.SD↗

arXiv2025

Spectral localization of single-nanoparticle plasmons through photonic substrate engineering

Surface plasmon resonances (SPRs) are crucial for confining light beyond the diffraction limit, yet heavy metal losses often limit their spectral localization. Here, we propose a practical strategy for enabling the spectral localization of single-nanoparticle SPRs through photonic substrate engineering, which creates distinct optical pathways (OPs) to tailor the electromagnetic environments around plasmonic nanoparticles. By analyzing the multiplication factor spectrum of the projected local density of states, we can trace and control these OPs, enabling strong spatial and spectral confinement of single-nanoparticle SPRs. Simulations reveal that a photonic crystal substrate can reduce the mode volume by fivefold and boost the quality factor by over 80 times compared to a metal nanoparticle on a dielectric substrate. Proof-of-concept experiments using two types of leaking Fabry-Perot photonic substrates demonstrate active manipulation of SPRs in both "open" and "closed" OP states. This multidimensional photonic substrate engineering establishes a customizable platform for single-nanoparticle plasmonics, potentially transforming applications that were previously limited by spectral localization.

physics.optics↗

arXiv2025

Mini-o3: Scaling Up Reasoning Patterns and Interaction Turns for Visual Search

Recent advances in large multimodal models have leveraged image-based tools with reinforcement learning to tackle visual problems. However, existing open-source approaches often exhibit monotonous reasoning patterns and allow only a limited number of interaction turns, making them inadequate for difficult tasks that require trial-and-error exploration. In this work, we address this limitation by scaling up tool-based interactions and introduce Mini-o3, a system that executes deep, multi-turn reasoning -- spanning tens of steps -- and achieves state-of-the-art performance on challenging visual search tasks. Our recipe for reproducing OpenAI o3-style behaviors comprises three key components. First, we construct the Visual Probe Dataset, a collection of thousands of challenging visual search problems designed for exploratory reasoning. Second, we develop an iterative data collection pipeline to obtain cold-start trajectories that exhibit diverse reasoning patterns, including depth-first search, trial-and-error, and goal maintenance. Third, we propose an over-turn masking strategy that prevents penalization of over-turn responses (those that hit the maximum number of turns) during reinforcement learning, thereby balancing training-time efficiency with test-time scalability. Despite training with an upper bound of only six interaction turns, our model generates trajectories that naturally scale to tens of turns at inference time, with accuracy improving as the number of turns increases. Extensive experiments demonstrate that Mini-o3 produces rich reasoning patterns and deep thinking paths, effectively solving challenging visual search problems.

cs.CV↗

arXiv2025

The enhanced X-ray Timing and Polarimetry mission -- eXTP for launch in 2030

In this paper we present the current status of the enhanced X-ray Timing and Polarimetry mission, which has been fully approved for launch in 2030. eXTP is a space science mission designed to study fundamental physics under extreme conditions of matter density, gravity, and magnetism. The mission aims at determining the equation of state of matter at supra-nuclear density, measuring the effects of quantum electro-dynamics, and understanding the dynamics of matter in strong-field gravity. In addition to investigating fundamental physics, the eXTP mission is poised to become a leading observatory for time-domain and multi-messenger astronomy in the 2030's, as well as providing observations of unprecedented quality on a variety of galactic and extragalactic objects. After briefly introducing the history and a summary of the scientific objectives of the eXTP mission, this paper presents a comprehensive overview of: 1) the cutting-edge technology, technical specifications, and anticipated performance of the mission's scientific instruments; 2) the full mission profile, encompassing spacecraft design, operational capabilities, and ground segment infrastructure.

astro-ph.HE↗

arXiv2025

FAST Observations of the Microstructure in Interpulse Pulsars

In this study, we investigate the microstructure properties of four pulsars (PSRs J0953+0755 (B0950+08), J0627+0706, J0826+2637 (B0823+26) and J1946+1805 (B1944+17)) using the Five-hundred-meter Aperture Spherical radio Telescope (FAST), with particular emphasis on identifying microstructure within interpulse (IP). Through the application of autocorrelation function (ACF) analysis and fast Fourier transform (FFT) techniques, we have systematically examined the periodicity of microstructure in these pulsars. Our findings represent the first successful detection of microstructure within IP. Furthermore, we conducted a comprehensive statistical analysis comparing the characteristic timescales ($τ_μ$) and the characteristic periods $P_μ$ of quasi-periodic microstructure between the main pulse (MP) and IP, and our results indicate that the $τ_μ$ and $P_μ$ of microstructure across components appear consistent within measurement errors for PSR J0627+0706, but microstructure in IP are relatively smaller than those in MP for PSR J0953+0755. Furthermore, the relationship between $P_μ$ of microstructure and the rotation period in neutron star populations was reconfirmed: $P_μ(\text{ms})=(1.337\pm0.114)\times P(\text{s})^{(1.063\pm0.038)}$.

astro-ph.HE↗

arXiv2025

Multi-modal Uncertainty Robust Tree Cover Segmentation For High-Resolution Remote Sensing Images

Recent advances in semantic segmentation of multi-modal remote sensing images have significantly improved the accuracy of tree cover mapping, supporting applications in urban planning, forest monitoring, and ecological assessment. Integrating data from multiple modalities-such as optical imagery, light detection and ranging (LiDAR), and synthetic aperture radar (SAR)-has shown superior performance over single-modality methods. However, these data are often acquired days or even months apart, during which various changes may occur, such as vegetation disturbances (e.g., logging, and wildfires) and variations in imaging quality. Such temporal misalignments introduce cross-modal uncertainty, especially in high-resolution imagery, which can severely degrade segmentation accuracy. To address this challenge, we propose MURTreeFormer, a novel multi-modal segmentation framework that mitigates and leverages aleatoric uncertainty for robust tree cover mapping. MURTreeFormer treats one modality as primary and others as auxiliary, explicitly modeling patch-level uncertainty in the auxiliary modalities via a probabilistic latent representation. Uncertain patches are identified and reconstructed from the primary modality's distribution through a VAE-based resampling mechanism, producing enhanced auxiliary features for fusion. In the decoder, a gradient magnitude attention (GMA) module and a lightweight refinement head (RH) are further integrated to guide attention toward tree-like structures and to preserve fine-grained spatial details. Extensive experiments on multi-modal datasets from Shanghai and Zurich demonstrate that MURTreeFormer significantly improves segmentation performance and effectively reduces the impact of temporally induced aleatoric uncertainty.

eess.IV↗

arXiv2025

A Survey: Learning Embodied Intelligence from Physical Simulators and World Models

The pursuit of artificial general intelligence (AGI) has placed embodied intelligence at the forefront of robotics research. Embodied intelligence focuses on agents capable of perceiving, reasoning, and acting within the physical world. Achieving robust embodied intelligence requires not only advanced perception and control, but also the ability to ground abstract cognition in real-world interactions. Two foundational technologies, physical simulators and world models, have emerged as critical enablers in this quest. Physical simulators provide controlled, high-fidelity environments for training and evaluating robotic agents, allowing safe and efficient development of complex behaviors. In contrast, world models empower robots with internal representations of their surroundings, enabling predictive planning and adaptive decision-making beyond direct sensory input. This survey systematically reviews recent advances in learning embodied AI through the integration of physical simulators and world models. We analyze their complementary roles in enhancing autonomy, adaptability, and generalization in intelligent robots, and discuss the interplay between external simulation and internal modeling in bridging the gap between simulated training and real-world deployment. By synthesizing current progress and identifying open challenges, this survey aims to provide a comprehensive perspective on the path toward more capable and generalizable embodied AI systems. We also maintain an active repository that contains up-to-date literature and open-source projects at https://github.com/NJU3DV-LoongGroup/Embodied-World-Models-Survey.

cs.RO↗

arXiv2025

FlashRecovery: Fast and Low-Cost Recovery from Failures for Large-Scale Training of LLMs

Large language models (LLMs) have made a profound impact across various fields due to their advanced capabilities. However, training these models at unprecedented scales requires extensive AI accelerator clusters and sophisticated parallelism strategies, which pose significant challenges in maintaining system reliability over prolonged training periods. A major concern is the substantial loss of training time caused by inevitable hardware and software failures. To address these challenges, we present FlashRecovery, a fast and low-cost failure recovery system comprising three core modules: (1) Active and real-time failure detection. This module performs continuous training state monitoring, enabling immediate identification of hardware and software failures within seconds, thus ensuring rapid incident response; (2) Scale-independent task restart. By employing different recovery strategies for normal and faulty nodes, combined with an optimized communication group reconstruction protocol, our approach ensures that the recovery time remains nearly constant, regardless of cluster scale; (3) Checkpoint-free recovery within one step. Our novel recovery mechanism enables single-step restoration, completely eliminating dependence on traditional checkpointing methods and their associated overhead. Collectively, these innovations enable FlashRecovery to achieve optimal Recovery Time Objective (RTO) and Recovery Point Objective (RPO), substantially improving the reliability and efficiency of long-duration LLM training. Experimental results demonstrate that FlashRecovery system can achieve training restoration on training cluster with 4, 800 devices in 150 seconds. We also verify that the time required for failure recovery is nearly consistent for different scales of training tasks.

cs.DC↗

arXiv2025

C-Flat++: Towards a More Efficient and Powerful Framework for Continual Learning

Balancing sensitivity to new tasks and stability for retaining past knowledge is crucial in continual learning (CL). Recently, sharpness-aware minimization has proven effective in transfer learning and has also been adopted in continual learning (CL) to improve memory retention and learning efficiency. However, relying on zeroth-order sharpness alone may favor sharper minima over flatter ones in certain settings, leading to less robust and potentially suboptimal solutions. In this paper, we propose \textbf{C}ontinual \textbf{Flat}ness (\textbf{C-Flat}), a method that promotes flatter loss landscapes tailored for CL. C-Flat offers plug-and-play compatibility, enabling easy integration with minimal modifications to the code pipeline. Besides, we present a general framework that integrates C-Flat into all major CL paradigms and conduct comprehensive comparisons with loss-minima optimizers and flat-minima-based CL methods. Our results show that C-Flat consistently improves performance across a wide range of settings. In addition, we introduce C-Flat++, an efficient yet effective framework that leverages selective flatness-driven promotion, significantly reducing the update cost required by C-Flat. Extensive experiments across multiple CL methods, datasets, and scenarios demonstrate the effectiveness and efficiency of our proposed approaches. Code is available at https://github.com/WanNaa/C-Flat.

cs.LG↗

arXiv2025

Modulation Instability-Induced Multimode Squeezing in Quadratic Frequency Combs

Lithium niobate (LN) microring resonators, characterized by an exceptionally high second-order nonlinear coefficient and superior electro-optic tunability, serve as an outstanding platform for the precise control of integrated quantum frequency combs (QFCs). In this study, we introduce a bipartite entanglement criterion to investigate the pairwise entanglement characteristics of QFCs generated via the spontaneous parametric down-conversion (SPDC) process in lithium niobate microring resonators operating below threshold. Furthermore, we propose a universal framework for analyzing multimode squeezing in quadratic frequency combs, enabling the realization of ultrabroadband and high-degree multimode squeezing. We further reveal the underlying physical mechanism: modulation instability (MI), regulated by temporal walk-off control, not only enables the formation of frequency combs but also induces multimode squeezing in the corresponding resonant modes. This study uncovers the previously unexplored role of on-chip multimode squeezing in quadratic frequency combs while facilitating collective noise suppression across multiple modes, thus holding substantial potential for advancing quantum precision measurement and quantum information processing.

quant-ph↗

arXiv2025

Describe, Don't Dictate: Semantic Image Editing with Natural Language Intent

Despite the progress in text-to-image generation, semantic image editing remains a challenge. Inversion-based algorithms unavoidably introduce reconstruction errors, while instruction-based models mainly suffer from limited dataset quality and scale. To address these problems, we propose a descriptive-prompt-based editing framework, named DescriptiveEdit. The core idea is to re-frame `instruction-based image editing' as `reference-image-based text-to-image generation', which preserves the generative power of well-trained Text-to-Image models without architectural modifications or inversion. Specifically, taking the reference image and a prompt as input, we introduce a Cross-Attentive UNet, which newly adds attention bridges to inject reference image features into the prompt-to-edit-image generation process. Owing to its text-to-image nature, DescriptiveEdit overcomes limitations in instruction dataset quality, integrates seamlessly with ControlNet, IP-Adapter, and other extensions, and is more scalable. Experiments on the Emu Edit benchmark show it improves editing accuracy and consistency.

cs.CV↗

arXiv2025

$H_\infty$ Performance Analysis for Almost Periodic Piecewise Linear Systems with Application to Roll-to-Roll Manufacturing Control

An almost periodic piecewise linear system (APPLS) is a type of piecewise linear system where the system cyclically switches between different modes, each with an uncertain but bounded dwell-time. Process regulation, especially disturbance rejection, is critical to the performance of these advanced systems. However, a method to guarantee disturbance rejection has not been developed. The objective of this study is to develop an $H_\infty$ performance analysis method for APPLSs, building on which an algorithm to synthesize practical $H_\infty$ controllers is proposed. As an application, the developed methods are demonstrated with an advanced manufacturing system -- roll-to-roll (R2R) dry transfer of two-dimensional materials and printed flexible electronics. Experimental results show that the proposed method enables a less conservative and much better performing $H_\infty$ controller compared with a baseline $H_\infty$ controller that does not account for the uncertain system switching structure.

eess.SY↗