arXiv subjects

Yiming Wang

Publications and source records attributed to Yiming Wang.

At least 127 records · Page 7Linked to original sources

arXiv2025

RetroDiff: Retrosynthesis as Multi-stage Distribution Interpolation

Retrosynthesis poses a key challenge in biopharmaceuticals, aiding chemists in finding appropriate reactant molecules for given product molecules. With reactants and products represented as 2D graphs, retrosynthesis constitutes a conditional graph-to-graph (G2G) generative task. Inspired by advancements in discrete diffusion models for graph generation, we aim to design a diffusion-based method to address this problem. However, integrating a diffusion-based G2G framework while retaining essential chemical reaction template information presents a notable challenge. Our key innovation involves a multi-stage diffusion process. We decompose the retrosynthesis procedure to first sample external groups from the dummy distribution given products, then generate external bonds to connect products and generated groups. Interestingly, this generation process mirrors the reverse of the widely adapted semi-template retrosynthesis workflow, \emph{i.e.} from reaction center identification to synthon completion. Based on these designs, we introduce Retrosynthesis Diffusion (RetroDiff), a novel diffusion-based method for the retrosynthesis task. Experimental results demonstrate that RetroDiff surpasses all semi-template methods in accuracy, and outperforms template-based and template-free methods in large-scale scenarios and molecular validity, respectively. Code: https://github.com/Alsace08/RetroDiff.

cs.LG↗

arXiv2025

Evaluating Attribute Confusion in Fashion Text-to-Image Generation

Despite the rapid advances in Text-to-Image (T2I) generation models, their evaluation remains challenging in domains like fashion, involving complex compositional generation. Recent automated T2I evaluation methods leverage pre-trained vision-language models to measure cross-modal alignment. However, our preliminary study reveals that they are still limited in assessing rich entity-attribute semantics, facing challenges in attribute confusion, i.e., when attributes are correctly depicted but associated to the wrong entities. To address this, we build on a Visual Question Answering (VQA) localization strategy targeting one single entity at a time across both visual and textual modalities. We propose a localized human evaluation protocol and introduce a novel automatic metric, Localized VQAScore (L-VQAScore), that combines visual localization with VQA probing both correct (reflection) and miss-localized (leakage) attribute generation. On a newly curated dataset featuring challenging compositional alignment scenarios, L-VQAScore outperforms state-of-the-art T2I evaluation methods in terms of correlation with human judgments, demonstrating its strength in capturing fine-grained entity-attribute associations. We believe L-VQAScore can be a reliable and scalable alternative to subjective evaluations.

cs.CV↗

arXiv2025

Suppression of shot noise at a Kondo destruction quantum critical point

Strange metal behavior has been observed in an expanding list of quantum materials, with heavy fermion metals serving as a prototype setting. Among the intriguing questions is the nature of charge carriers; there is an increasing recognition that the quasiparticles are lost, as captured by Kondo destruction quantum criticality. Among the recent experimental advances is the measurement of shot noise in a heavy-fermion strange metal. We are thus motivated to study current fluctuations by advancing a minimal Bose-Fermi Kondo lattice model, which admits a well-defined large-$N$ limit. Showing that the model in equilibrium captures the essential physics of Kondo destruction, we proceed to derive quantum kinetic equations and compute shot noise to the leading nontrivial order in $1/N$. Our results reveal a strong suppression of the shot noise at the Kondo destruction quantum critical point, thereby providing the understanding of the striking experiment. Broader implications of our results are discussed.

cond-mat.str-el↗

arXiv2025

Distributed Lyapunov Functions for Nonlinear Networks

Nonlinear networks are often multistable, exhibiting coexisting stable states with competing regions of attraction (ROAs). As a result, ROAs can have complex "tentacle-like" morphologies that are challenging to characterize analytically or computationally. In addition, the high dimensionality of the state space prohibits the automated construction of Lyapunov functions using state-of-the-art optimization methods, such as sum-of-squares (SOS) programming. In this letter, we propose a distributed approach for the construction of Lyapunov functions based solely on local information. To this end, we establish an augmented comparison lemma that characterizes the existence conditions of partial Lyapunov functions, while also accounting for residual effects caused by the associated dimensionality reduction. These theoretical results allow us to formulate an SOS optimization that iteratively constructs such partial functions, whose aggregation forms a composite Lyapunov function. The resulting composite function provides accurate convex approximations of both the volumes and shapes of the ROAs. We validate our method on networks of van der Pol and Ising oscillators, demonstrating its effectiveness in characterizing high-dimensional systems with non-convex ROAs.

eess.SY↗

arXiv2025

Multimodal Fusion SLAM with Fourier Attention

Visual SLAM is particularly challenging in environments affected by noise, varying lighting conditions, and darkness. Learning-based optical flow algorithms can leverage multiple modalities to address these challenges, but traditional optical flow-based visual SLAM approaches often require significant computational resources.To overcome this limitation, we propose FMF-SLAM, an efficient multimodal fusion SLAM method that utilizes fast Fourier transform (FFT) to enhance the algorithm efficiency. Specifically, we introduce a novel Fourier-based self-attention and cross-attention mechanism to extract features from RGB and depth signals. We further enhance the interaction of multimodal features by incorporating multi-scale knowledge distillation across modalities. We also demonstrate the practical feasibility of FMF-SLAM in real-world scenarios with real time performance by integrating it with a security robot by fusing with a global positioning module GNSS-RTK and global Bundle Adjustment. Our approach is validated using video sequences from TUM, TartanAir, and our real-world datasets, showcasing state-of-the-art performance under noisy, varying lighting, and dark conditions.Our code and datasets are available at https://github.com/youjie-zhou/FMF-SLAM.git.

cs.CV↗

arXiv2025

STAMImputer: Spatio-Temporal Attention MoE for Traffic Data Imputation

Traffic data imputation is fundamentally important to support various applications in intelligent transportation systems such as traffic flow prediction. However, existing time-to-space sequential methods often fail to effectively extract features in block-wise missing data scenarios. Meanwhile, the static graph structure for spatial feature propagation significantly constrains the models flexibility in handling the distribution shift issue for the nonstationary traffic data. To address these issues, this paper proposes a SpatioTemporal Attention Mixture of experts network named STAMImputer for traffic data imputation. Specifically, we introduce a Mixture of Experts (MoE) framework to capture latent spatio-temporal features and their influence weights, effectively imputing block missing. A novel Low-rank guided Sampling Graph ATtention (LrSGAT) mechanism is designed to dynamically balance the local and global correlations across road networks. The sampled attention vectors are utilized to generate dynamic graphs that capture real-time spatial correlations. Extensive experiments are conducted on four traffic datasets for evaluation. The result shows STAMImputer achieves significantly performance improvement compared with existing SOTA approaches. Our codes are available at https://github.com/RingBDStack/STAMImupter.

cs.LG↗

arXiv2025

Nematicity in iron pnictides: phase competition and emergent symmetry

The phase diagram of iron-based superconductors contains a host of electronic orders, which are intimately connected with their superconductivity. Here we analyze the fluctuations of one type of nematic order in another. Our analysis leads to an emergent U(1) symmetry at a first-order transition between a nematic phase and a $C_4$-symmetric charge-ordered phase. We characterize the continuous symmetry in terms of a certain hidden Lie algebra that links the different orders. This emergent symmetry leads to a Goldstone mode at the transition and causes softening of excitations in the nematic and charge sectors near the transition. The underlying physics bears a resemblance to the anisotropic XZ spin model, with the nematic order and charge $C_4$ order parameters playing the roles of the $x$ and $z$ components of the magnetization vector, respectively. We provide the experimental evidence in support of the proposed effects, and discuss the general implications of our results for the physics of iron-based superconductors and other correlated systems.

cond-mat.str-el↗

arXiv2025

Decoding Speaker-Normalized Pitch from EEG for Mandarin Perception

The same speech content produced by different speakers exhibits significant differences in pitch contour, yet listeners' semantic perception remains unaffected. This phenomenon may stem from the brain's perception of pitch contours being independent of individual speakers' pitch ranges. In this work, we recorded electroencephalogram (EEG) while participants listened to Mandarin monosyllables with varying tones, phonemes, and speakers. The CE-ViViT model is proposed to decode raw or speaker-normalized pitch contours directly from EEG. Experimental results demonstrate that the proposed model can decode pitch contours with modest errors, achieving performance comparable to state-of-the-art EEG regression methods. Moreover, speaker-normalized pitch contours were decoded more accurately, supporting the neural encoding of relative pitch.

cs.SD↗

arXiv2025

Fully nonequilibrium Hall response from Berry curvature

In topological materials, Berry curvature leads to intrinsic Hall responses. Focusing on time-reversal symmetric systems with broken inversion symmetry, a spontaneoous (zero magnetic field) Hall effect is expected to develop under an applied electric field. Motivated by recent developments in Weyl-Kondo semimetals, here we advance a fully nonequilibrium (FNE) Hall response due to the Berry curvature. In particular, we show that, while the spontaneous Hall current is quadratic in the previously described regime of weak electric field, due to the contribution from the dipole moment of the Berry curvature, the FNE Hall response for non-perturbative electric fields is not controlled by the Berry curvature dipole. Remarkably, the FNE Hall response resembles what happens in systems that break the microscopic time-reversal symmetry. We illustrate the universality of these results by comparing them with their counterparts in systems with any higher-multipole of the Berry curvature. The implications of our results for the understanding of strongly correlated topological semimetals are discussed.

cond-mat.mes-hall↗

arXiv2025

Nonlinear Model Predictive Control for Leaderless UAV Formation Flying with Collision Avoidance under Directed Graphs

This paper studies the leaderless formation flying problem with collision avoidance for a group of unmanned aerial vehicles (UAVs), which requires the UAVs to navigate through cluttered environments without colliding while maintaining the formation. The communication network among the UAVs is structured as a directed graph that includes a directed spanning tree. A novel distributed nonlinear model predictive control (NMPC) method based on the model reference adaptive consensus (MRACon) framework is proposed. Within this framework, each UAV tracks an assigned reference output generated by a linear reference model that utilizes relative measurements as input. Subsequently, the NMPC method penalizes the tracking error between the output of the reference model and that of the actual model while also establishing constraint sets for collision avoidance and physical limitations to achieve distributed and safe formation control. Finally, simulations and hardware experiments are conducted to verify the effectiveness of the proposed method.

eess.SY↗

arXiv2025

Seeing the Abstract: Translating the Abstract Language for Vision Language Models

Natural language goes beyond dryly describing visual content. It contains rich abstract concepts to express feeling, creativity and properties that cannot be directly perceived. Yet, current research in Vision Language Models (VLMs) has not shed light on abstract-oriented language. Our research breaks new ground by uncovering its wide presence and under-estimated value, with extensive analysis. Particularly, we focus our investigation on the fashion domain, a highly-representative field with abstract expressions. By analyzing recent large-scale multimodal fashion datasets, we find that abstract terms have a dominant presence, rivaling the concrete ones, providing novel information, and being useful in the retrieval task. However, a critical challenge emerges: current general-purpose or fashion-specific VLMs are pre-trained with databases that lack sufficient abstract words in their text corpora, thus hindering their ability to effectively represent abstract-oriented language. We propose a training-free and model-agnostic method, Abstract-to-Concrete Translator (ACT), to shift abstract representations towards well-represented concrete ones in the VLM latent space, using pre-trained models and existing multimodal databases. On the text-to-image retrieval task, despite being training-free, ACT outperforms the fine-tuned VLMs in both same- and cross-dataset settings, exhibiting its effectiveness with a strong generalization capability. Moreover, the improvement introduced by ACT is consistent with various VLMs, making it a plug-and-play solution.

cs.CV↗

arXiv2025

Achievable Second-Order Asymptotics for MAC and RAC with Additive Non-Gaussian Noise

We first study the two-user additive noise multiple access channel (MAC) where the noise distribution is arbitrary. For such a MAC, we use spherical codebooks and either joint nearest neighbor (JNN) or successive interference cancellation (SIC) decoding. Under both decoding methods, we derive second-order achievable rate regions and compare the finite blocklength performance between JNN and SIC decoding. Our results indicate that although the first-order rate regions of JNN and SIC decoding are identical, JNN decoding has better second-order asymptotic performance. When specialized to the Gaussian noise, we provide an alternative achievability proof to the result by MolavianJazi and Laneman (T-IT, 2015). Furthermore, we generalize our results to the random access channel (RAC) where neither the transmitters nor the receiver knows the user activity pattern. We use spherical-type codebooks and a rateless transmission scheme combining JNN/SIC decoding, and derive second-order achievability bounds. Comparing second-order achievability results of JNN and SIC decoding in a RAC, we show that JNN decoding achieves strictly larger first-order asymptotic rate. When specialized to Gaussian noise, our second-order asymptotic results recover the corresponding results of Yavas, Kostina, and Effros (T-IT, 2021) up to second-order.

cs.IT↗

arXiv2025

Two-dimensional non-van der Waals niobium nitride nanosheets with high-temperature two-gap superconductivity

The exploration of the superconductivity in two-dimensional materials has garnered significant attention due to their promising low-power applications and fundamental scientific interest. Here, we report some novel stable non-van der Waals Nb$_x$N$_{x+1}$ ($x$ = 1-4) monolayers derived from the NbN bulk exfoliated along the (001) plane, as identified through first-principles calculations. Among these monolayers, Nb$_2$N$_3$, which crystallizes in the $P \overline{6} m2$ symmetry, stands out with an exceptional superconducting transition temperature of 77.8 K, setting a new high-$T_c$ benchmark for two-dimensional transition metal nitrides and binary compounds. Our detailed analysis reveals that the strong superconductivity in Nb$_2$N$_3$ is driven by phonon modes dominated by N vibrations, with significant electron-phonon coupling contributions from N-$p$ and Nb-$d$ electronic states. Using the anisotropic Migdal-Eliashberg framework, we further determine the two-gap nature of the superconductivity in the Nb$_2$N$_3$ monolayer, characterized by pronounced electron-phonon coupling and anisotropic energy gaps. These results advance our understanding of superconductivity in 2D transition metal nitride and highlight their potential for nanoscale superconducting applications.

cond-mat.supr-con↗

arXiv2025

Bernstein Bounds for Caustics

Systematically simulating specular light transport requires an exhaustive search for primitive tuples containing admissible paths. Given the extreme inefficiency of enumerating all combinations, we propose to significantly reduce the search domain by sampling such tuples. The challenge is to design proper sampling probabilities that keep the noise level controllable. Our key insight is that by bounding the range of irradiance contributed by each primitive tuple at a given position, we can sample a subset of primitive tuples with potentially high contributions. Although low-contribution tuples are assigned a negligible probability, the overall variance remains low. Therefore, we derive vertex position and irradiance bounds for each primitive tuple, introducing a bounding property of rational functions on the Bernstein basis. When formulating position and irradiance expressions into rational functions, we handle non-rational components through remainder variables to maintain validity. Finally, we carefully design the sampling probabilities by optimizing the upper bound of the variance, expressed only using the position and irradiance bound. The proposed primitive sampling is intrinsically unbiased. It can be seamlessly combined with various unbiased and biased root-finding techniques within a local primitive domain. Extensive evaluations show that our method enables fast and reliable rendering of complex caustic effects.

cs.GR↗

arXiv2025

Wan: Open and Advanced Large-Scale Video Generative Models

This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.

cs.CV↗

arXiv2025

Error analysis for temporal second-order finite element approximations of axisymmetric mean curvature flow of genus-1 surfaces

Existing studies on the convergence of numerical methods for curvature flows primarily focus on first-order temporal schemes. In this paper, we establish a novel error analysis for parametric finite element approximations of genus-1 axisymmetric mean curvature flow, formulated using two classical second-order time-stepping methods: the Crank-Nicolson method and the BDF2 method. Our results establish optimal error bounds in both the L^2-norm and H^1-norm, along with a superconvergence result in the H^1-norm for each fully discrete approximation. Finally, we perform convergence experiments to validate the theoretical findings and present numerical simulations for various genus-1 surfaces. Through a series of comparative experiments, we also demonstrate that the methods proposed in this paper exhibit significant mesh advantages.

math.NA↗

arXiv2025

PerLA: Perceptive 3D Language Assistant

Enabling Large Language Models (LLMs) to understand the 3D physical world is an emerging yet challenging research direction. Current strategies for processing point clouds typically downsample the scene or divide it into smaller parts for separate analysis. However, both approaches risk losing key local details or global contextual information. In this paper, we introduce PerLA, a 3D language assistant designed to be more perceptive to both details and context, making visual representations more informative for the LLM. PerLA captures high-resolution (local) details in parallel from different point cloud areas and integrates them with (global) context obtained from a lower-resolution whole point cloud. We present a novel algorithm that preserves point cloud locality through the Hilbert curve and effectively aggregates local-to-global information via cross-attention and a graph neural network. Lastly, we introduce a novel loss for local representation consensus to promote training stability. PerLA outperforms state-of-the-art 3D language assistants, with gains of up to +1.34 CiDEr on ScanQA for question answering, and +4.22 on ScanRefer and +3.88 on Nr3D for dense captioning. https://gfmei.github.io/PerLA/

cs.CV↗

arXiv2025

Vocabulary-Free 3D Instance Segmentation with Vision and Language Assistant

Most recent 3D instance segmentation methods are open vocabulary, offering a greater flexibility than closed-vocabulary methods. Yet, they are limited to reasoning within a specific set of concepts, \ie the vocabulary, prompted by the user at test time. In essence, these models cannot reason in an open-ended fashion, i.e., answering "List the objects in the scene.''. We introduce the first method to address 3D instance segmentation in a setting that is void of any vocabulary prior, namely a vocabulary-free setting. We leverage a large vision-language assistant and an open-vocabulary 2D instance segmenter to discover and ground semantic categories on the posed images. To form 3D instance mask, we first partition the input point cloud into dense superpoints, which are then merged into 3D instance masks. We propose a novel superpoint merging strategy via spectral clustering, accounting for both mask coherence and semantic coherence that are estimated from the 2D object instance masks. We evaluate our method using ScanNet200 and Replica, outperforming existing methods in both vocabulary-free and open-vocabulary settings. Code will be made available. Project page: https://gfmei.github.io/PoVo

cs.CV↗