arXiv subjects

Lin Zhang

Publications and source records attributed to Lin Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2025

INR-Bench: A Unified Benchmark for Implicit Neural Representations in Multi-Domain Regression and Reconstruction

Implicit Neural Representations (INRs) have gained success in various signal processing tasks due to their advantages of continuity and infinite resolution. However, the factors influencing their effectiveness and limitations remain underexplored. To better understand these factors, we leverage insights from Neural Tangent Kernel (NTK) theory to analyze how model architectures (classic MLP and emerging KAN), positional encoding, and nonlinear primitives affect the response to signals of varying frequencies. Building on this analysis, we introduce INR-Bench, the first comprehensive benchmark specifically designed for multimodal INR tasks. It includes 56 variants of Coordinate-MLP models (featuring 4 types of positional encoding and 14 activation functions) and 22 Coordinate-KAN models with distinct basis functions, evaluated across 9 implicit multimodal tasks. These tasks cover both forward and inverse problems, offering a robust platform to highlight the strengths and limitations of different neural models, thereby establishing a solid foundation for future research. The code and dataset are available at https://github.com/lif314/INR-Bench.

cs.LG↗

arXiv2025

Operator entanglement in $\mathrm{SU}(2)$-symmetric dissipative quantum many-body dynamics

The presence of symmetries can lead to nontrivial dynamics of operator entanglement in open quantum many-body systems, which characterizes the cost of an matrix product density operator (MPDO) representation of the density matrix in the tensor-network methods and provides a measure for the corresponding classical simulability. One example is the $\mathrm{U}(1)$-symmetric open quantum systems with dephasing, in which the operator entanglement increases logarithmically at late times instead of being suppressed by the dephasing. Here we numerically study the far-from-equilibrium dynamics of operator entanglement in a dissipative quantum many-body system with the more complicated $\mathrm{SU}(2)$ symmetry and dissipations beyond dephasing. We show that after the initial rise and fall, the operator entanglement also increases again in a logarithmic manner at late times in the $\mathrm{SU}(2)$-symmetric case. We find that this behavior can be fully understood from the corresponding $\mathrm{U}(1)$ subsymmetry by considering the symmetry-resolved operator entanglement. But unlike the $\mathrm{U}(1)$-symmetric case with dephasing, both the classical Shannon entropy associated with the probabilities for the half system being in different symmetry sectors and the corresponding symmetry-resolved operator entanglement have nontrivial contributions to the late time logarithmic growth of operator entanglement. Our results show evidence that the logarithmic growth of operator entanglement at long times is a generic behavior of dissipative quantum many-body dynamics with $\mathrm{U}(1)$ as the symmetry or subsymmetry and for more broad dissipations beyond dephasing. By breaking the $\mathrm{SU}(2)$ symmetry of our quantum many-body dynamics to $\mathrm{U}(1)$, we also show that the latter property is valid even for open quantum systems with only $\mathrm{U}(1)$ symmetry.

quant-ph↗

arXiv2025

Informative Text-Image Alignment for Visual Affordance Learning with Foundation Models

Visual affordance learning is crucial for robots to understand and interact effectively with the physical world. Recent advances in this field attempt to leverage pre-trained knowledge of vision-language foundation models to learn affordance properties with limited training data, providing a novel paradigm for visual affordance learning. However, these methods overlook the significance of maintaining feature alignment between visual images and language descriptions for identifying affordance areas with textual guidance, and thus may lead to suboptimal results. In this paper, we present an informative framework for text-guided affordance learning, which involves information-based constraints to achieve text-image alignment at feature level. Specifically, we design an affordance mutual information constraint that helps learn appropriate textual prompts and task-oriented visual features simultaneously by maximizing the mutual information between the features of the affordance areas in the input images and the corresponding textual prompts. In addition, we propose an object-level information constraint that maximizes the mutual information between the visual features of a given object and the text features of the category it belongs to. This enables the model to capture high-quality representations for the object, providing more reliable semantic priors for identifying affordance regions. Experimental results on the AGD20K dataset show that the proposed method outperforms existing approaches and achieves the new state-of-the-art in one-shot affordance learning.

cs.CV↗

arXiv2025

Sequential Token Merging: Revisiting Hidden States

Vision Mambas (ViMs) achieve remarkable success with sub-quadratic complexity, but their efficiency remains constrained by quadratic token scaling with image resolution. While existing methods address token redundancy, they overlook ViMs' intrinsic Limited Directional Sequential Dependence (LDSD) - a critical information flow mechanism revealed in our analysis. We further identify Mamba's selective scan enables gradual information aggregation in hidden states. Based on these insights, we propose Sequential Token Merging (STM), featuring: 1) Bidirectional nearest neighbor merging to preserve sequential dependencies through symmetric spatial aggregation, and 2) Hidden states protection to stabilize the hidden states around the class token. STM strategically leverages Mamba's layer-wise loss convergence to convert temporal forgetfulness into stability. Experiments demonstrate STM's superiority: 1.0% accuracy drop for ViM-Ti at 20% token reduction, and only 1.4% degradation for ViM-S at 40% reduction. Our method achieves state-of-the-art efficiency with minimal complexity, while providing new insights into state-space model dynamics. Codes will be released soon.

cs.CV↗

arXiv2025

Harnessing Uncertainty: Entropy-Modulated Policy Gradients for Long-Horizon LLM Agents

In long-horizon tasks, recent agents based on Large Language Models (LLMs) face a significant challenge that sparse, outcome-based rewards make it difficult to assign credit to intermediate steps. Previous methods mainly focus on creating dense reward signals to guide learning, either through traditional reinforcement learning techniques like inverse reinforcement learning or by using Process Reward Models for step-by-step feedback. In this paper, we identify a fundamental problem in the learning dynamics of LLMs: the magnitude of policy gradients is inherently coupled with the entropy, which leads to inefficient small updates for confident correct actions and potentially destabilizes large updates for uncertain ones. To resolve this, we propose Entropy-Modulated Policy Gradients (EMPG), a framework that re-calibrates the learning signal based on step-wise uncertainty and the final task outcome. EMPG amplifies updates for confident correct actions, penalizes confident errors, and attenuates updates from uncertain steps to stabilize exploration. We further introduce a bonus term for future clarity that encourages agents to find more predictable solution paths. Through comprehensive experiments on three challenging agent tasks, WebShop, ALFWorld, and Deep Search, we demonstrate that EMPG achieves substantial performance gains and significantly outperforms strong policy gradient baselines. Project page is at https://empgseed-seed.github.io/

cs.LG↗

arXiv2025

Multiple solutions to a class of $p$-Laplacian Schrödinger equations

In this paper, we will prove the existence of infinitely many solutions to the following equation by utilizing the variational perturbation method \begin{equation*} -div(A(x,u)|\nabla u|^{p-2}\nabla u)+\frac{1}{p}A_{t}(x,u)|\nabla u|^{p}+V(x)|u|^{p-2}u=g(x, u), ~~ x\in\mathbb{R}^{N}, \end{equation*} where $2<p<N$, $V(x)$ represents a coercive potential. It should be emphasized that the multiplicity of solutions is unsuitable to be discussed under the framework of $W^{1,p}(\mathbb{R}^{N})\cap L^{\infty}(\mathbb{R}^{N})$, as the functional corresponding to the above equation does not satisfy the Palais-Smale condition in this space. To this end, we will use the variational perturbation method to construct novel perturbation space and perturbation functional. By relying on the established conclusions regarding the multiplicity of solutions for classical $p$-Laplacian equations, we shall analyze the equation in question. Our result addresses positively the open question from Candela et. al in CVPDE.

math.AP↗

arXiv2025

Nonradial normalized solutions for a quasilinear Schrödinger equation via dual approach

In this paper, we will utilize the dual method to construct multiple nonradial normalized solutions of the following quasilinear Schrödinger equation: \begin{equation*} -Δu-Δ(|u|^{2})u-μu=|u|^{p-2}u, \qquad in \quad \mathbb{R}^N, \end{equation*} subject to a mass-subcritical constraint. It should be emphasized that the nonradial result of this equation is new. Besides that, when considering the nonradial problem, it is necessary to construct a new workspace to ensure the compactness. Meanwhile, in this paper, we will expand on the method mentioned in TMNA.

math.AP↗

arXiv2025

Revealing the anisotropic charge-density-wave order of TiSe$_2$ through high harmonic generation

Titanium diselenide (TiSe$_{2}$) is a transition-metal dichalcogenide material that undergoes a charge-density-wave (CDW) transition at $T_{c}\approx 200\,\mathrm{K}$. In a recent experiment [I. Tyulnev {\it et al.}, Commun. Mater. 6, 152 (2025)], the high harmonic generation (HHG) spectra of this material has been studied, which exhibits asymmetric behavior with respect to the polarization angle of the incident light and provides a new perspective to the CDW phase transition. In this work, we work out a theoretical explanation for the experimentally observed behavior of HHG spectra. We propose a simplified phenomenological mean-field model for this material, based on which the HHG spectra is calculated through the time-dependent Schr{\" o}dinger equation. This model correctly describes the measured intensity distribution of the third-, fifth-, and seventh-order harmonic generation as a function of polarization direction and reveals a strong asymmetry due to the anisotropic CDW order in the low temperature phase. Our work provides a basis for applying high harmonic spectroscopy to reveal a new perspective on the nature of CDW orders.

cond-mat.str-el↗

arXiv2025

M3HG: Multimodal, Multi-scale, and Multi-type Node Heterogeneous Graph for Emotion Cause Triplet Extraction in Conversations

Emotion Cause Triplet Extraction in Multimodal Conversations (MECTEC) has recently gained significant attention in social media analysis, aiming to extract emotion utterances, cause utterances, and emotion categories simultaneously. However, the scarcity of related datasets, with only one published dataset featuring highly uniform dialogue scenarios, hinders model development in this field. To address this, we introduce MECAD, the first multimodal, multi-scenario MECTEC dataset, comprising 989 conversations from 56 TV series spanning a wide range of dialogue contexts. In addition, existing MECTEC methods fail to explicitly model emotional and causal contexts and neglect the fusion of semantic information at different levels, leading to performance degradation. In this paper, we propose M3HG, a novel model that explicitly captures emotional and causal contexts and effectively fuses contextual information at both inter- and intra-utterance levels via a multimodal heterogeneous graph. Extensive experiments demonstrate the effectiveness of M3HG compared with existing state-of-the-art methods. The codes and dataset are available at https://github.com/redifinition/M3HG.

cs.CL↗

arXiv2025

Gorgeous: Revisiting the Data Layout for Disk-Resident High-Dimensional Vector Search

Similarity-based vector search underpins many important applications, but a key challenge is processing massive vector datasets (e.g., in TBs). To reduce costs, some systems utilize SSDs as the primary data storage. They employ a proximity graph, which connects similar vectors to form a graph and is the state-of-the-art index for vector search. However, these systems are hindered by sub-optimal data layouts that fail to effectively utilize valuable memory space to reduce disk access and suffer from poor locality for accessing disk-resident data. Through extensive profiling and analysis, we found that the structure of the proximity graph index is accessed more frequently than the vectors themselves, yet existing systems do not distinguish between the two. To address this problem, we design the Gorgeous system with the principle of prioritizing graph structure over vectors. Specifically, Gorgeous features a memory cache that keeps the adjacency lists of graph nodes to improve cache hits and a disk block format that explicitly stores neighbors' adjacency lists along with a vector to enhance data locality. Experimental results show that Gorgeous consistently outperforms two state-of-the-art disk-based systems for vector search, boosting average query throughput by over 60% and reducing query latency by over 35%.

cs.DB↗

arXiv2025

Towards Generalized Source Tracing for Codec-Based Deepfake Speech

Recent attempts at source tracing for codec-based deepfake speech (CodecFake), generated by neural audio codec-based speech generation (CoSG) models, have exhibited suboptimal performance. However, how to train source tracing models using simulated CoSG data while maintaining strong performance on real CoSG-generated audio remains an open challenge. In this paper, we show that models trained solely on codec-resynthesized data tend to overfit to non-speech regions and struggle to generalize to unseen content. To mitigate these challenges, we introduce the Semantic-Acoustic Source Tracing Network (SASTNet), which jointly leverages Whisper for semantic feature encoding and Wav2vec2 with AudioMAE for acoustic feature encoding. Our proposed SASTNet achieves state-of-the-art performance on the CoSG test set of the CodecFake+ dataset, demonstrating its effectiveness for reliable source tracing.

cs.SD↗

arXiv2025

HierMoE: Accelerating MoE Training with Hierarchical Token Deduplication and Expert Swap

The sparsely activated mixture-of-experts (MoE) transformer has become a common architecture for large language models (LLMs) due to its sparsity, which requires fewer computational demands while easily scaling the model size. In MoE models, each MoE layer requires to dynamically choose tokens to activate particular experts for computation while the activated experts may not be located in the same device or GPU as the token. However, this leads to substantial communication and load imbalances across all GPUs, which obstructs the scalability of distributed systems within a GPU cluster. To this end, we introduce HierMoE to accelerate the training of MoE models by two topology-aware techniques: 1) token deduplication to reduce the communication traffic, and 2) expert swap to balance the workloads among all GPUs. To enable the above two proposed approaches to be more general, we build theoretical models aimed at achieving the best token duplication and expert swap strategy under different model configurations and hardware environments. We implement our prototype HierMoE system atop Megatron-LM and conduct experiments on a 32-GPU cluster with DeepSeek-V3 and Qwen3-30B-A3B models. Experimental results show that our HierMoE achieves $1.55\times$ to $3.32\times$ faster communication and delivers $1.18\times$ to $1.27\times$ faster end-to-end training compared to state-of-the-art MoE training systems, Tutel-2DH, SmartMoE, and Megatron-LM.

cs.DC↗

arXiv2025

SC-Captioner: Improving Image Captioning with Self-Correction by Reinforcement Learning

We propose SC-Captioner, a reinforcement learning framework that enables the self-correcting capability of image caption models. Our crucial technique lies in the design of the reward function to incentivize accurate caption corrections. Specifically, the predicted and reference captions are decomposed into object, attribute, and relation sets using scene-graph parsing algorithms. We calculate the set difference between sets of initial and self-corrected captions to identify added and removed elements. These elements are matched against the reference sets to calculate correctness bonuses for accurate refinements and mistake punishments for wrong additions and removals, thereby forming the final reward. For image caption quality assessment, we propose a set of metrics refined from CAPTURE that alleviate its incomplete precision evaluation and inefficient relation matching problems. Furthermore, we collect a fine-grained annotated image caption dataset, RefinedCaps, consisting of 6.5K diverse images from COCO dataset. Experiments show that applying SC-Captioner on large visual-language models can generate better image captions across various scenarios, significantly outperforming the direct preference optimization training strategy.

cs.CV↗

arXiv2025

Scaling Up Audio-Synchronized Visual Animation: An Efficient Training Paradigm

Recent advances in audio-synchronized visual animation enable control of video content using audios from specific classes. However, existing methods rely heavily on expensive manual curation of high-quality, class-specific training videos, posing challenges to scaling up to diverse audio-video classes in the open world. In this work, we propose an efficient two-stage training paradigm to scale up audio-synchronized visual animation using abundant but noisy videos. In stage one, we automatically curate large-scale videos for pretraining, allowing the model to learn diverse but imperfect audio-video alignments. In stage two, we finetune the model on manually curated high-quality examples, but only at a small scale, significantly reducing the required human effort. We further enhance synchronization by allowing each frame to access rich audio context via multi-feature conditioning and window attention. To efficiently train the model, we leverage pretrained text-to-video generator and audio encoders, introducing only 1.9\% additional trainable parameters to learn audio-conditioning capability without compromising the generator's prior knowledge. For evaluation, we introduce AVSync48, a benchmark with videos from 48 classes, which is 3$\times$ more diverse than previous benchmarks. Extensive experiments show that our method significantly reduces reliance on manual curation by over 10$\times$, while generalizing to many open classes.

cs.CV↗

arXiv2025

Localizing Audio-Visual Deepfakes via Hierarchical Boundary Modeling

Audio-visual temporal deepfake localization under the content-driven partial manipulation remains a highly challenging task. In this scenario, the deepfake regions are usually only spanning a few frames, with the majority of the rest remaining identical to the original. To tackle this, we propose a Hierarchical Boundary Modeling Network (HBMNet), which includes three modules: an Audio-Visual Feature Encoder that extracts discriminative frame-level representations, a Coarse Proposal Generator that predicts candidate boundary regions, and a Fine-grained Probabilities Generator that refines these proposals using bidirectional boundary-content probabilities. From the modality perspective, we enhance audio-visual learning through dedicated encoding and fusion, reinforced by frame-level supervision to boost discriminability. From the temporal perspective, HBMNet integrates multi-scale cues and bidirectional boundary-content relationships. Experiments show that encoding and fusion primarily improve precision, while frame-level supervision boosts recall. Each module (audio-visual fusion, temporal scales, bi-directionality) contributes complementary benefits, collectively enhancing localization performance. HBMNet outperforms BA-TFD and UMMAFormer and shows improved potential scalability with more training data.

cs.SD↗

arXiv2025

Codec-Based Deepfake Source Tracing via Neural Audio Codec Taxonomy

Recent advances in neural audio codec-based speech generation (CoSG) models have produced remarkably realistic audio deepfakes. We refer to deepfake speech generated by CoSG systems as codec-based deepfake, or CodecFake. Although existing anti-spoofing research on CodecFake predominantly focuses on verifying the authenticity of audio samples, almost no attention was given to tracing the CoSG used in generating these deepfakes. In CodecFake generation, processes such as speech-to-unit encoding, discrete unit modeling, and unit-to-speech decoding are fundamentally based on neural audio codecs. Motivated by this, we introduce source tracing for CodecFake via neural audio codec taxonomy, which dissects neural audio codecs to trace CoSG. Our experimental results on the CodecFake+ dataset provide promising initial evidence for the feasibility of CodecFake source tracing while also highlighting several challenges that warrant further investigation.

cs.SD↗

arXiv2025

Learning from Rendering: Realistic and Controllable Extreme Rainy Image Synthesis for Autonomous Driving Simulation

Autonomous driving simulators provide an effective and low-cost alternative for evaluating or enhancing visual perception models. However, the reliability of evaluation depends on the diversity and realism of the generated scenes. Extreme weather conditions, particularly extreme rainfalls, are rare and costly to capture in real-world settings. While simulated environments can help address this limitation, existing rainy image synthesizers often suffer from poor controllability over illumination and limited realism, which significantly undermines the effectiveness of the model evaluation. To that end, we propose a learning-from-rendering rainy image synthesizer, which combines the benefits of the realism of rendering-based methods and the controllability of learning-based methods. To validate the effectiveness of our extreme rainy image synthesizer on semantic segmentation task, we require a continuous set of well-labeled extreme rainy images. By integrating the proposed synthesizer with the CARLA driving simulator, we develop CARLARain an extreme rainy street scene simulator which can obtain paired rainy-clean images and labels under complex illumination conditions. Qualitative and quantitative experiments validate that CARLARain can effectively improve the accuracy of semantic segmentation models in extreme rainy scenes, with the models' accuracy (mIoU) improved by 5% - 8% on the synthetic dataset and significantly enhanced in real extreme rainy scenarios under complex illuminations. Our source code and datasets are available at https://github.com/kb824999404/CARLARain/.

cs.CV↗

arXiv2025

Differential Contrastive Training for Gaze Estimation

The complex application scenarios have raised critical requirements for precise and generalizable gaze estimation methods. Recently, the pre-trained CLIP has achieved remarkable performance on various vision tasks, but its potentials have not been fully exploited in gaze estimation. In this paper, we propose a novel Differential Contrastive Training strategy, which boosts gaze estimation performance with the help of the CLIP. Accordingly, a Differential Contrastive Gaze Estimation network (DCGaze) composed of a Visual Appearance-aware branch and a Semantic Differential-aware branch is introduced. The Visual Appearance-aware branch is essentially a primary gaze estimation network and it incorporates an Adaptive Feature-refinement Unit (AFU) and a Double-head Gaze Regressor (DGR), which both help the primary network to extract informative and gaze-related appearance features. Moreover, the Semantic Difference-aware branch is designed on the basis of the CLIP's text encoder to reveal the semantic difference of gazes. This branch could further empower the Visual Appearance-aware branch with the capability of characterizing the gaze-related semantic information. Extensive experimental results on four challenging datasets over within and cross-domain tasks demonstrate the effectiveness of our DCGaze.The code is available at https://github.com/LinZhang-bjtu/DCGaze.

cs.CV↗