arXiv subjects

Jun Liu

Publications and source records attributed to Jun Liu.

At least 343 records · Page 19Linked to original sources

arXiv2025

PromptMID: Modal Invariant Descriptors Based on Diffusion and Vision Foundation Models for Optical-SAR Image Matching

The ideal goal of image matching is to achieve stable and efficient performance in unseen domains. However, many existing learning-based optical-SAR image matching methods, despite their effectiveness in specific scenarios, exhibit limited generalization and struggle to adapt to practical applications. Repeatedly training or fine-tuning matching models to address domain differences is not only not elegant enough but also introduces additional computational overhead and data production costs. In recent years, general foundation models have shown great potential for enhancing generalization. However, the disparity in visual domains between natural and remote sensing images poses challenges for their direct application. Therefore, effectively leveraging foundation models to improve the generalization of optical-SAR image matching remains challenge. To address the above challenges, we propose PromptMID, a novel approach that constructs modality-invariant descriptors using text prompts based on land use classification as priors information for optical and SAR image matching. PromptMID extracts multi-scale modality-invariant features by leveraging pre-trained diffusion models and visual foundation models (VFMs), while specially designed feature aggregation modules effectively fuse features across different granularities. Extensive experiments on optical-SAR image datasets from four diverse regions demonstrate that PromptMID outperforms state-of-the-art matching methods, achieving superior results in both seen and unseen domains and exhibiting strong cross-domain generalization capabilities. The source code will be made publicly available https://github.com/HanNieWHU/PromptMID.

cs.CV↗

arXiv2025

FedBM: Stealing Knowledge from Pre-trained Language Models for Heterogeneous Federated Learning

Federated learning (FL) has shown great potential in medical image computing since it provides a decentralized learning paradigm that allows multiple clients to train a model collaboratively without privacy leakage. However, current studies have shown that data heterogeneity incurs local learning bias in classifiers and feature extractors of client models during local training, leading to the performance degradation of a federation system. To address these issues, we propose a novel framework called Federated Bias eliMinating (FedBM) to get rid of local learning bias in heterogeneous federated learning (FL), which mainly consists of two modules, i.e., Linguistic Knowledge-based Classifier Construction (LKCC) and Concept-guided Global Distribution Estimation (CGDE). Specifically, LKCC exploits class concepts, prompts and pre-trained language models (PLMs) to obtain concept embeddings. These embeddings are used to estimate the latent concept distribution of each class in the linguistic space. Based on the theoretical derivation, we can rely on these distributions to pre-construct a high-quality classifier for clients to achieve classification optimization, which is frozen to avoid classifier bias during local training. CGDE samples probabilistic concept embeddings from the latent concept distributions to learn a conditional generator to capture the input space of the global model. Three regularization terms are introduced to improve the quality and utility of the generator. The generator is shared by all clients and produces pseudo data to calibrate updates of local feature extractors. Extensive comparison experiments and ablation studies on public datasets demonstrate the superior performance of FedBM over state-of-the-arts and confirm the effectiveness of each module, respectively. The code is available at https://github.com/CUHK-AIM-Group/FedBM.

cs.CV↗

arXiv2025

Baichuan-Audio: A Unified Framework for End-to-End Speech Interaction

We introduce Baichuan-Audio, an end-to-end audio large language model that seamlessly integrates audio understanding and generation. It features a text-guided aligned speech generation mechanism, enabling real-time speech interaction with both comprehension and generation capabilities. Baichuan-Audio leverages a pre-trained ASR model, followed by multi-codebook discretization of speech at a frame rate of 12.5 Hz. This multi-codebook setup ensures that speech tokens retain both semantic and acoustic information. To further enhance modeling, an independent audio head is employed to process audio tokens, effectively capturing their unique characteristics. To mitigate the loss of intelligence during pre-training and preserve the original capabilities of the LLM, we propose a two-stage pre-training strategy that maintains language understanding while enhancing audio modeling. Following alignment, the model excels in real-time speech-based conversation and exhibits outstanding question-answering capabilities, demonstrating its versatility and efficiency. The proposed model demonstrates superior performance in real-time spoken dialogue and exhibits strong question-answering abilities. Our code, model and training data are available at https://github.com/baichuan-inc/Baichuan-Audio

cs.CL↗

arXiv2025

Unified Prompt Attack Against Text-to-Image Generation Models

Text-to-Image (T2I) models have advanced significantly, but their growing popularity raises security concerns due to their potential to generate harmful images. To address these issues, we propose UPAM, a novel framework to evaluate the robustness of T2I models from an attack perspective. Unlike prior methods that focus solely on textual defenses, UPAM unifies the attack on both textual and visual defenses. Additionally, it enables gradient-based optimization, overcoming reliance on enumeration for improved efficiency and effectiveness. To handle cases where T2I models block image outputs due to defenses, we introduce Sphere-Probing Learning (SPL) to enable optimization even without image results. Following SPL, our model bypasses defenses, inducing the generation of harmful content. To ensure semantic alignment with attacker intent, we propose Semantic-Enhancing Learning (SEL) for precise semantic control. UPAM also prioritizes the naturalness of adversarial prompts using In-context Naturalness Enhancement (INE), making them harder for human examiners to detect. Additionally, we address the issue of iterative queries--common in prior methods and easily detectable by API defenders--by introducing Transferable Attack Learning (TAL), allowing effective attacks with minimal queries. Extensive experiments validate UPAM's superiority in effectiveness, efficiency, naturalness, and low query detection rates.

cs.CV↗

arXiv2025

Some sets of first category in product Calderón-Lozanovski\uı spaces on hypergroups

Let $K$ be a locally compact hypergroup with a left Haar measure $μ$ and $Ω$ be a Banach ideal of $μ$-measurable complex-valued functions on $K$. For Young functions $\{φ_i\}_{i=1,2,3}$, let $Ω_{φ_i}(K)$ be the corresponding Calderón--Lozanovski\uı space associated with $φ_i$ on $K$. Motivated by the remarkable work of Akbarbaglu et al. in [Adv. Math. 312 (2017), 737-763], in this article, the authors give several sufficient conditions for the sets $$\left\{(f,g)\inΩ_{φ_1}(K)\timesΩ_{φ_2}(K):\ |f|\ast |g|\inΩ_{φ_3}(K)\right\}$$ and $$\left\{(f,g)\inΩ_{φ_1}(K)\timesΩ_{φ_2}(K):\ \exists\,x\in U,\ (|f|\ast |g|)(x)<\infty\right\}$$ to be of first category in the sense of Baire, where $U\subset K$ denotes a compact set. All these results are new even for Orlicz(-Lorentz) spaces on hypergroups.

math.CA↗

arXiv2025

ZeroPS: High-quality Cross-modal Knowledge Transfer for Zero-Shot 3D Part Segmentation

Zero-shot 3D part segmentation is a challenging and fundamental task. In this work, we propose a novel pipeline, ZeroPS, which achieves high-quality knowledge transfer from 2D pretrained foundation models (FMs), SAM and GLIP, to 3D object point clouds. We aim to explore the natural relationship between multi-view correspondence and the FMs' prompt mechanism and build bridges on it. In ZeroPS, the relationship manifests as follows: 1) lifting 2D to 3D by leveraging co-viewed regions and SAM's prompt mechanism, 2) relating 1D classes to 3D parts by leveraging 2D-3D view projection and GLIP's prompt mechanism, and 3) enhancing prediction performance by leveraging multi-view observations. Extensive evaluations on the PartNetE and AKBSeg benchmarks demonstrate that ZeroPS significantly outperforms the SOTA method across zero-shot unlabeled and instance segmentation tasks. ZeroPS does not require additional training or fine-tuning for the FMs. ZeroPS applies to both simulated and real-world data. It is hardly affected by domain shift. The project page is available at https://luis2088.github.io/ZeroPS_page.

cs.CV↗

arXiv2025

Achieving electrode smoothing by controlling the nucleation phase of metal deposition through polymer-substrate binding

Polymer additives [like polyethylene oxide (PEO)] are widely used for smooth electrode deposition in aqueous zinc and a number of other battery systems currently investigated for energy storage applications. However, the precise mechanism by which they regulate morphology and suppress dendrite formation remains unclear. In this study, we address this knowledge gap by using in-situ electrochemical atomic force microscopy (EC-AFM) to directly observe the interfacial evolution during Zn electrodeposition and polymer adsorption on copper (Cu) substrates in the presence of varying concentrations of ZnSO4 and PEO. Contrary to previous literature assumptions which emphasize the binding to the growing Zn crystal surfaces or Zn2+ ions, our results demonstrate that PEO smooths Zn films by promoting nucleation of (002)-oriented Zn platelets through interactions with the Cu substrate. Density functional theory (DFT) simulations support this finding by showing that PEO adsorption on Cu modifies the interfacial energy of Zn/Cu/electrolyte interfaces, favoring the stabilization of Zn (002) on the Cu substrate, as well as confines Zn electrodeposition to a narrow near-surface region. These findings elucidate a novel design principle for electrode smoothing, emphasizing the importance of substrate selection paired with polymer additives that exhibit an attractive interaction with the substrate, but minimal interaction with growing crystals, offering a mechanistic perspective for improved battery performance.

cond-mat.mtrl-sci↗

arXiv2025

Quasi-perfect spatiotemporal optical vortex with suppressed mode degradation

Spatiotemporal optical vortex (STOV) carrying transverse orbital angular momentum (OAM) enriches the family of vortex beams and exhibit unique properties. Typically, a high-order STOV with an intensity null degrades into multiple first-order STOVs embedded within a single wave packet during propagation, a phenomenon known as time diffraction or mode degradation. However, this degradation limits the applicability of STOVs in specialized fields. Therefore, the generation of mode degradation-suppressed STOVs (MDS-STOVs) is of significant for both practical applications and theoretical studies. Herein, we theoretically analyze the generation of MDS-STOVs by utilizing a conical phase to localize the energy of the STOV into a ring-shaped structure. For MDS-STOVs with large topological charges (TCs), the ring-shaped profile can be well-maintained, and the rapid expansion of the beam size with increasing TC is significantly suppressed compared to conventional STOVs. As a result, these MDS-STOVs can be regarded as quasi-perfect STOVs (QPSTOVs). Furthermore, QPSTOVs exhibit strong resistance to group delay dispersion (GDD), eliminating the need for precise dispersion control and facilitating their generation and application. This work advances our understanding of the physical properties of light carrying transverse OAM and opens up exciting avenues for the application of STOVs in diverse fields, such as optical communication and quantum information processing.

physics.optics↗

arXiv2025

HyLiFormer: Hyperbolic Linear Attention for Skeleton-based Human Action Recognition

Transformers have demonstrated remarkable performance in skeleton-based human action recognition, yet their quadratic computational complexity remains a bottleneck for real-world applications. To mitigate this, linear attention mechanisms have been explored but struggle to capture the hierarchical structure of skeleton data. Meanwhile, the Poincaré model, as a typical hyperbolic geometry, offers a powerful framework for modeling hierarchical structures but lacks well-defined operations for existing mainstream linear attention. In this paper, we propose HyLiFormer, a novel hyperbolic linear attention Transformer tailored for skeleton-based action recognition. Our approach incorporates a Hyperbolic Transformation with Curvatures (HTC) module to map skeleton data into hyperbolic space and a Hyperbolic Linear Attention (HLA) module for efficient long-range dependency modeling. Theoretical analysis and extensive experiments on NTU RGB+D and NTU RGB+D 120 datasets demonstrate that HyLiFormer significantly reduces computational complexity while preserving model accuracy, making it a promising solution for efficiency-critical applications.

cs.CV↗

arXiv2025

Binding of the three-hadron $DD^{*}K$ system from the lattice effective field theory

We employ the nuclear lattice effective field theory (NLEFT), an efficient tool for nuclear ab initio calculations, to solve the asymmetric multihadron systems. We take the $DD^*K$ three-body system as an illustration to demonstrate the capability of the method. Here the two-body chiral interactions between $D$, $D^*$, and $K$ are regulated with a soft lattice regulator and calibrated with the binding energies of the $T_{cc}^+$, $D^{*}_{s0}(2317)$, and $D_{s1}(2460)$ molecular states. We then calculate the three-body binding energy using the NLEFT and analyze the systematic uncertainties due to the finite volume effects, the sliding cutoff, and the leading-order three-body forces. Even when the three-body interaction is repulsive (even as large as the infinite repulsive interaction), the three-body system has a bound state unambiguously with binding energy no larger than the $D_{s1}(2460)D$ threshold. To check the renormalization group invariance of our framework, we extract the first excited state. We find that when the ground state is fixed, the first excited states with various cutoffs coincide with each other when the cubic size goes larger. In addition, the standard angular momentum and parity projection technique is implemented for the quantum numbers of the ground and excited states. We find that both of them are $S$-wave states with quantum number $J^{P}=1^-$. Because the three-body state contains two charm quarks, it is easier to be detected in the Large Hadron Collider.

hep-ph↗

arXiv2025

Conterfactual Generative Zero-Shot Semantic Segmentation

zero-shot learning is an essential part of computer vision. As a classical downstream task, zero-shot semantic segmentation has been studied because of its applicant value. One of the popular zero-shot semantic segmentation methods is based on the generative model Most new proposed works added structures on the same architecture to enhance this model. However, we found that, from the view of causal inference, the result of the original model has been influenced by spurious statistical relationships. Thus the performance of the prediction shows severe bias. In this work, we consider counterfactual methods to avoid the confounder in the original model. Based on this method, we proposed a new framework for zero-shot semantic segmentation. Our model is compared with baseline models on two real-world datasets, Pascal-VOC and Pascal-Context. The experiment results show proposed models can surpass previous confounded models and can still make use of additional structures to improve the performance. We also design a simple structure based on Graph Convolutional Networks (GCN) in this work.

cs.CV↗

arXiv2025

Not Every Patch is Needed: Towards a More Efficient and Effective Backbone for Video-based Person Re-identification

This paper proposes a new effective and efficient plug-and-play backbone for video-based person re-identification (ReID). Conventional video-based ReID methods typically use CNN or transformer backbones to extract deep features for every position in every sampled video frame. Here, we argue that this exhaustive feature extraction could be unnecessary, since we find that different frames in a ReID video often exhibit small differences and contain many similar regions due to the relatively slight movements of human beings. Inspired by this, a more selective, efficient paradigm is explored in this paper. Specifically, we introduce a patch selection mechanism to reduce computational cost by choosing only the crucial and non-repetitive patches for feature extraction. Additionally, we present a novel network structure that generates and utilizes pseudo frame global context to address the issue of incomplete views resulting from sparse inputs. By incorporating these new designs, our backbone can achieve both high performance and low computational cost. Extensive experiments on multiple datasets show that our approach reduces the computational cost by 74\% compared to ViT-B and 28\% compared to ResNet50, while the accuracy is on par with ViT-B and outperforms ResNet50 significantly.

cs.CV↗

arXiv2025

Baichuan-Omni-1.5 Technical Report

We introduce Baichuan-Omni-1.5, an omni-modal model that not only has omni-modal understanding capabilities but also provides end-to-end audio generation capabilities. To achieve fluent and high-quality interaction across modalities without compromising the capabilities of any modality, we prioritized optimizing three key aspects. First, we establish a comprehensive data cleaning and synthesis pipeline for multimodal data, obtaining about 500B high-quality data (text, audio, and vision). Second, an audio-tokenizer (Baichuan-Audio-Tokenizer) has been designed to capture both semantic and acoustic information from audio, enabling seamless integration and enhanced compatibility with MLLM. Lastly, we designed a multi-stage training strategy that progressively integrates multimodal alignment and multitask fine-tuning, ensuring effective synergy across all modalities. Baichuan-Omni-1.5 leads contemporary models (including GPT4o-mini and MiniCPM-o 2.6) in terms of comprehensive omni-modal capabilities. Notably, it achieves results comparable to leading models such as Qwen2-VL-72B across various multimodal medical benchmarks.

cs.CL↗

arXiv2025

ResoFilter: Fine-grained Synthetic Data Filtering for Large Language Models through Data-Parameter Resonance Analysis

Large language models (LLMs) have shown remarkable effectiveness across various domains, with data augmentation methods utilizing GPT for synthetic data generation becoming prevalent. However, the quality and utility of augmented data remain questionable, and current methods lack clear metrics for evaluating data characteristics. To address these challenges, we propose ResoFilter, a novel method that integrates models, data, and tasks to refine datasets. ResoFilter leverages the fine-tuning process to obtain Data-Parameter features for data selection, offering improved interpretability by representing data characteristics through model weights. Our experiments demonstrate that ResoFilter achieves comparable results to full-scale fine-tuning using only half the data in mathematical tasks and exhibits strong generalization across different models and domains. This method provides valuable insights for constructing synthetic datasets and evaluating high-quality data, offering a promising solution for enhancing data augmentation techniques and improving training dataset quality for LLMs. For reproducibility, we will release our code and data upon acceptance.

cs.CL↗

arXiv2025

AI-Generated Content (AIGC) for Various Data Modalities: A Survey

AI-generated content (AIGC) methods aim to produce text, images, videos, 3D assets, and other media using AI algorithms. Due to its wide range of applications and the potential of recent works, AIGC developments -- especially in Machine Learning (ML) and Deep Learning (DL) -- have been attracting significant attention, and this survey focuses on comprehensively reviewing such advancements in ML/DL. AIGC methods have been developed for various data modalities, such as image, video, text, 3D shape, 3D scene, 3D human avatar, 3D motion, and audio -- each presenting unique characteristics and challenges. Furthermore, there have been significant developments in cross-modality AIGC methods, where generative methods receive conditioning input in one modality and produce outputs in another. Examples include going from various modalities to image, video, 3D, and audio. This paper provides a comprehensive review of AIGC methods across different data modalities, including both single-modality and cross-modality methods, highlighting the various challenges, representative works, and recent technical directions in each setting. We also survey the representative datasets throughout the modalities, and present comparative results for various modalities. Moreover, we discuss the typical applications of AIGC methods in various domains, challenges, and future research directions.

cs.CV↗

arXiv2025

EvoChart: A Benchmark and a Self-Training Approach Towards Real-World Chart Understanding

Chart understanding enables automated data analysis for humans, which requires models to achieve highly accurate visual comprehension. While existing Visual Language Models (VLMs) have shown progress in chart understanding, the lack of high-quality training data and comprehensive evaluation benchmarks hinders VLM chart comprehension. In this paper, we introduce EvoChart, a novel self-training method for generating synthetic chart data to enhance VLMs' capabilities in real-world chart comprehension. We also propose EvoChart-QA, a noval benchmark for measuring models' chart comprehension abilities in real-world scenarios. Specifically, EvoChart is a unique self-training data synthesis approach that simultaneously produces high-quality training corpus and a high-performance chart understanding model. EvoChart-QA consists of 650 distinct real-world charts collected from 140 different websites and 1,250 expert-curated questions that focus on chart understanding. Experimental results on various open-source and proprietary VLMs tested on EvoChart-QA demonstrate that even the best proprietary model, GPT-4o, achieves only 49.8% accuracy. Moreover, the EvoChart method significantly boosts the performance of open-source VLMs on real-world chart understanding tasks, achieving 54.2% accuracy on EvoChart-QA.

cs.CV↗

arXiv2025

RoRA: Efficient Fine-Tuning of LLM with Reliability Optimization for Rank Adaptation

Fine-tuning helps large language models (LLM) recover degraded information and enhance task performance. Although Low-Rank Adaptation (LoRA) is widely used and effective for fine-tuning, we have observed that its scaling factor can limit or even reduce performance as the rank size increases. To address this issue, we propose RoRA (Rank-adaptive Reliability Optimization), a simple yet effective method for optimizing LoRA's scaling factor. By replacing $α/r$ with $α/\sqrt{r}$, RoRA ensures improved performance as rank size increases. Moreover, RoRA enhances low-rank adaptation in fine-tuning uncompressed models and excels in the more challenging task of accuracy recovery when fine-tuning pruned models. Extensive experiments demonstrate the effectiveness of RoRA in fine-tuning both uncompressed and pruned models. RoRA surpasses the state-of-the-art (SOTA) in average accuracy and robustness on LLaMA-7B/13B, LLaMA2-7B, and LLaMA3-8B, specifically outperforming LoRA and DoRA by 6.5% and 2.9% on LLaMA-7B, respectively. In pruned model fine-tuning, RoRA shows significant advantages; for SHEARED-LLAMA-1.3, a LLaMA-7B with 81.4% pruning, RoRA achieves 5.7% higher average accuracy than LoRA and 3.9% higher than DoRA.

cs.LG↗

arXiv2025

Physics-Informed Neural Network Lyapunov Functions: PDE Characterization, Learning, and Verification

We provide a systematic investigation of using physics-informed neural networks to compute Lyapunov functions. We encode Lyapunov conditions as a partial differential equation (PDE) and use this for training neural network Lyapunov functions. We analyze the analytical properties of the solutions to the Lyapunov and Zubov PDEs. In particular, we show that employing the Zubov equation in training neural Lyapunov functions can lead to approximate regions of attraction close to the true domain of attraction. We also examine approximation errors and the convergence of neural approximations to the unique solution of Zubov's equation. We then provide sufficient conditions for the learned neural Lyapunov functions that can be readily verified by satisfiability modulo theories (SMT) solvers, enabling formal verification of both local stability analysis and region-of-attraction estimates in the large. Through a number of nonlinear examples, ranging from low to high dimensions, we demonstrate that the proposed framework can outperform traditional sums-of-squares (SOS) Lyapunov functions obtained using semidefinite programming (SDP).

math.OC↗