arXiv subjects

Yujia Wu

Publications and source records attributed to Yujia Wu.

34 records · Page 2Linked to original sources

arXiv2025

MM-R1: Unleashing the Power of Unified Multimodal Large Language Models for Personalized Image Generation

Multimodal Large Language Models (MLLMs) with unified architectures excel across a wide range of vision-language tasks, yet aligning them with personalized image generation remains a significant challenge. Existing methods for MLLMs are frequently subject-specific, demanding a data-intensive fine-tuning process for every new subject, which limits their scalability. In this paper, we introduce MM-R1, a framework that integrates a cross-modal Chain-of-Thought (X-CoT) reasoning strategy to unlock the inherent potential of unified MLLMs for personalized image generation. Specifically, we structure personalization as an integrated visual reasoning and generation process: (1) grounding subject concepts by interpreting and understanding user-provided images and contextual cues, and (2) generating personalized images conditioned on both the extracted subject representations and user prompts. To further enhance the reasoning capability, we adopt Grouped Reward Proximal Policy Optimization (GRPO) to explicitly align the generation. Experiments demonstrate that MM-R1 unleashes the personalization capability of unified MLLMs to generate images with high subject fidelity and strong text alignment in a zero-shot manner.

cs.CV↗

arXiv2025

Vision-G1: Towards General Vision Language Reasoning with Multi-Domain Data Curation

Despite their success, current training pipelines for reasoning VLMs focus on a limited range of tasks, such as mathematical and logical reasoning. As a result, these models face difficulties in generalizing their reasoning capabilities to a wide range of domains, primarily due to the scarcity of readily available and verifiable reward data beyond these narrowly defined areas. Moreover, integrating data from multiple domains is challenging, as the compatibility between domain-specific datasets remains uncertain. To address these limitations, we build a comprehensive RL-ready visual reasoning dataset from 46 data sources across 8 dimensions, covering a wide range of tasks such as infographic, mathematical, spatial, cross-image, graphic user interface, medical, common sense and general science. We propose an influence function based data selection and difficulty based filtering strategy to identify high-quality training samples from this dataset. Subsequently, we train the VLM, referred to as Vision-G1, using multi-round RL with a data curriculum to iteratively improve its visual reasoning capabilities. Our model achieves state-of-the-art performance across various visual reasoning benchmarks, outperforming similar-sized VLMs and even proprietary models like GPT-4o and Gemini-1.5 Flash. The model, code and dataset are publicly available at https://github.com/yuh-zha/Vision-G1.

cs.CV↗

arXiv2025

Matchings in matroids over abelian groups, II

The concept of matchings originated in group theory to address a linear algebra problem related to canonical forms for symmetric tensors. In an abelian group $(G,+)$, a matching is a bijection $f: A \to B$ between two finite subsets $A$ and $B$ of $G$ such that $a + f(a) \notin A$ for all $a \in A$. A group $G$ has the matching property if, for every two finite subsets $A, B \subset G$ of the same size with $0 \notin B$, there exists a matching from $A$ to $B$. In prior work [5], matroid analogues of results concerning matchings in groups were introduced and established. This paper serves as a sequel, extending that line of inquiry by investigating sparse paving, panhandle, and Schubert matroids through the lens of matchability. While some proofs draw upon earlier findings on the matchability of sparse paving matroids, the paper is designed to be self-contained and accessible without reference to the preceding sequel. Our approach combines tools from both matroid theory and additive number theory.

math.CO↗

arXiv2025

Qwen-Image Technical Report

We present Qwen-Image, an image generation foundation model in the Qwen series that achieves significant advances in complex text rendering and precise image editing. To address the challenges of complex text rendering, we design a comprehensive data pipeline that includes large-scale data collection, filtering, annotation, synthesis, and balancing. Moreover, we adopt a progressive training strategy that starts with non-text-to-text rendering, evolves from simple to complex textual inputs, and gradually scales up to paragraph-level descriptions. This curriculum learning approach substantially enhances the model's native text rendering capabilities. As a result, Qwen-Image not only performs exceptionally well in alphabetic languages such as English, but also achieves remarkable progress on more challenging logographic languages like Chinese. To enhance image editing consistency, we introduce an improved multi-task training paradigm that incorporates not only traditional text-to-image (T2I) and text-image-to-image (TI2I) tasks but also image-to-image (I2I) reconstruction, effectively aligning the latent representations between Qwen2.5-VL and MMDiT. Furthermore, we separately feed the original image into Qwen2.5-VL and the VAE encoder to obtain semantic and reconstructive representations, respectively. This dual-encoding mechanism enables the editing module to strike a balance between preserving semantic consistency and maintaining visual fidelity. Qwen-Image achieves state-of-the-art performance, demonstrating its strong capabilities in both image generation and editing across multiple benchmarks.

cs.CV↗

arXiv2025

A Goodness-of-Fit Test for Sparse Networks

The stochastic block model (SBM) has been widely used to analyze network data. Various goodness-of-fit tests have been proposed to assess the adequacy of model structures. To the best of our knowledge, however, none of the existing approaches are applicable for sparse networks in which the connection probability of any two communities is of order log(n)/n, and the number of communities is divergent. To fill this gap, we propose a novel goodness-of-fit test for the stochastic block model. The key idea is to construct statistics by sampling the maximum entry-deviations of the adjacency matrix that the negative impacts of network sparsity are alleviated by the sampling process. We demonstrate theoretically that the proposed test statistic converges to the Type-I extreme value distribution under the null hypothesis regardless of the network structure. Accordingly, it can be applied to both dense and sparse networks. In addition, we obtain the asymptotic power against alternatives. Moreover, we introduce a bootstrap-corrected test statistic to improve the finite sample performance, recommend an augmented test statistic to increase the power, and extend the proposed test to the degree-corrected SBM. Simulation studies and two empirical examples with both dense and sparse networks indicate that the proposed method performs well.

stat.ME↗

arXiv2025

On Path to Multimodal Historical Reasoning: HistBench and HistAgent

Recent advances in large language models (LLMs) have led to remarkable progress across domains, yet their capabilities in the humanities, particularly history, remain underexplored. Historical reasoning poses unique challenges for AI, involving multimodal source interpretation, temporal inference, and cross-linguistic analysis. While general-purpose agents perform well on many existing benchmarks, they lack the domain-specific expertise required to engage with historical materials and questions. To address this gap, we introduce HistBench, a new benchmark of 414 high-quality questions designed to evaluate AI's capacity for historical reasoning and authored by more than 40 expert contributors. The tasks span a wide range of historical problems-from factual retrieval based on primary sources to interpretive analysis of manuscripts and images, to interdisciplinary challenges involving archaeology, linguistics, or cultural history. Furthermore, the benchmark dataset spans 29 ancient and modern languages and covers a wide range of historical periods and world regions. Finding the poor performance of LLMs and other agents on HistBench, we further present HistAgent, a history-specific agent equipped with carefully designed tools for OCR, translation, archival search, and image understanding in History. On HistBench, HistAgent based on GPT-4o achieves an accuracy of 27.54% pass@1 and 36.47% pass@2, significantly outperforming LLMs with online search and generalist agents, including GPT-4o (18.60%), DeepSeek-R1(14.49%) and Open Deep Research-smolagents(20.29% pass@1 and 25.12% pass@2). These results highlight the limitations of existing LLMs and generalist agents and demonstrate the advantages of HistAgent for historical reasoning.

cs.AI↗

arXiv2025

Observation of Near-Critical Kibble-Zurek Scaling in Rydberg Atom Arrays

The Kibble-Zurek scaling reveals the universal dynamics when a system is linearly ramped across a symmetry-breaking phase transition. However, in reality, inevitable finite-size effects or symmetrybreaking perturbations can often smear out the critical point and render the phase transition into a smooth crossover. In this letter, we show experimentally that the precise Kibble-Zurek scaling can be retained in the near-critical crossover regime, not necessarily crossing the critical point strictly. The key ingredient to achieving this near-critical Kibble-Zurek scaling is that the system size and the symmetry-breaking field must be appropriately scaled following the variation of ramping speeds. The experiment is performed in a reconfigurable Rydberg atom array platform, where the Rydberg blockade effect induces a Z2 symmetry-breaking transition. The atom array platform enables precise control of the system size and the zigzag geometry as a symmetry-breaking field. Therefore, we can demonstrate notable differences in the precision of the Kibble-Zurek scaling with or without properly scaling the system size and the zigzag geometry. Our results strengthen the Kibble-Zurek scaling as an increasingly valuable tool for investigating phase transition in quantum simulation platforms.

cond-mat.quant-gas↗

arXiv2025

Tensor-Fused Multi-View Graph Contrastive Learning

Graph contrastive learning (GCL) has emerged as a promising approach to enhance graph neural networks' (GNNs) ability to learn rich representations from unlabeled graph-structured data. However, current GCL models face challenges with computational demands and limited feature utilization, often relying only on basic graph properties like node degrees and edge attributes. This constrains their capacity to fully capture the complex topological characteristics of real-world phenomena represented by graphs. To address these limitations, we propose Tensor-Fused Multi-View Graph Contrastive Learning (TensorMV-GCL), a novel framework that integrates extended persistent homology (EPH) with GCL representations and facilitates multi-scale feature extraction. Our approach uniquely employs tensor aggregation and compression to fuse information from graph and topological features obtained from multiple augmented views of the same graph. By incorporating tensor concatenation and contraction modules, we reduce computational overhead by separating feature tensor aggregation and transformation. Furthermore, we enhance the quality of learned topological features and model robustness through noise-injected EPH. Experiments on molecular, bioinformatic, and social network datasets demonstrate TensorMV-GCL's superiority, outperforming 15 state-of-the-art methods in graph classification tasks across 9 out of 11 benchmarks while achieving comparable results on the remaining two. The code for this paper is publicly available at https://github.com/CS-SAIL/Tensor-MV-GCL.git.

cs.LG↗

arXiv2024

Precise Facial Landmark Detection by Dynamic Semantic Aggregation Transformer

At present, deep neural network methods have played a dominant role in face alignment field. However, they generally use predefined network structures to predict landmarks, which tends to learn general features and leads to mediocre performance, e.g., they perform well on neutral samples but struggle with faces exhibiting large poses or occlusions. Moreover, they cannot effectively deal with semantic gaps and ambiguities among features at different scales, which may hinder them from learning efficient features. To address the above issues, in this paper, we propose a Dynamic Semantic-Aggregation Transformer (DSAT) for more discriminative and representative feature (i.e., specialized feature) learning. Specifically, a Dynamic Semantic-Aware (DSA) model is first proposed to partition samples into subsets and activate the specific pathways for them by estimating the semantic correlations of feature channels, making it possible to learn specialized features from each subset. Then, a novel Dynamic Semantic Specialization (DSS) model is designed to mine the homogeneous information from features at different scales for eliminating the semantic gap and ambiguities and enhancing the representation ability. Finally, by integrating the DSA model and DSS model into our proposed DSAT in both dynamic architecture and dynamic parameter manners, more specialized features can be learned for achieving more precise face alignment. It is interesting to show that harder samples can be handled by activating more feature channels. Extensive experiments on popular face alignment datasets demonstrate that our proposed DSAT outperforms state-of-the-art models in the literature.Our code is available at https://github.com/GERMINO-LiuHe/DSAT.

cs.CV↗

arXiv2024

DiffLoRA: Generating Personalized Low-Rank Adaptation Weights with Diffusion

Personalized text-to-image generation has gained significant attention for its capability to generate high-fidelity portraits of specific identities conditioned on user-defined prompts. Existing methods typically involve test-time fine-tuning or incorporating an additional pre-trained branch. However, these approaches struggle to simultaneously address efficiency, identity fidelity, and the preservation of the model's original generative capabilities. In this paper, we propose DiffLoRA, an efficient method that leverages the diffusion model as a hypernetwork to predict personalized Low-Rank Adaptation (LoRA) weights based on the reference images. By incorporating these LoRA weights into the off-the-shelf text-to-image model, DiffLoRA enables zero-shot personalization during inference, eliminating the need for post-processing optimization. Moreover, we introduce a novel identity-oriented LoRA weights construction pipeline to facilitate the training process of DiffLoRA. The dataset generated through this pipeline enables DiffLoRA to produce consistently high-quality LoRA weights. Notably, the distinctive properties of the diffusion model enhance the generation of superior weights by employing probabilistic modeling to capture intricate structural patterns and thoroughly explore the weight space. Comprehensive experimental results demonstrate that DiffLoRA outperforms existing personalization approaches across multiple benchmarks, achieving both time efficiency and maintaining identity fidelity throughout the personalization process.

cs.CV↗

arXiv2024

Conditional Prediction ROC Bands for Graph Classification

Graph classification in medical imaging and drug discovery requires accuracy and robust uncertainty quantification. To address this need, we introduce Conditional Prediction ROC (CP-ROC) bands, offering uncertainty quantification for ROC curves and robustness to distributional shifts in test data. Although developed for Tensorized Graph Neural Networks (TGNNs), CP-ROC is adaptable to general Graph Neural Networks (GNNs) and other machine learning models. We establish statistically guaranteed coverage for CP-ROC under a local exchangeability condition. This addresses uncertainty challenges for ROC curves under non-iid setting, ensuring reliability when test graph distributions differ from training data. Empirically, to establish local exchangeability for TGNNs, we introduce a data-driven approach to construct local calibration sets for graphs. Comprehensive evaluations show that CP-ROC significantly improves prediction reliability across diverse tasks. This method enhances uncertainty quantification efficiency and reliability for ROC curves, proving valuable for real-world applications with non-iid objects.

cs.LG↗

arXiv2024

Conditional Uncertainty Quantification for Tensorized Topological Neural Networks

Graph Neural Networks (GNNs) have become the de facto standard for analyzing graph-structured data, leveraging message-passing techniques to capture both structural and node feature information. However, recent studies have raised concerns about the statistical reliability of uncertainty estimates produced by GNNs. This paper addresses this crucial challenge by introducing a novel technique for quantifying uncertainty in non-exchangeable graph-structured data, while simultaneously reducing the size of label prediction sets in graph classification tasks. We propose Conformalized Tensor-based Topological Neural Networks (CF-T2NN), a new approach for rigorous prediction inference over graphs. CF-T2NN employs tensor decomposition and topological knowledge learning to navigate and interpret the inherent uncertainty in decision-making processes. This method enables a more nuanced understanding and handling of prediction uncertainties, enhancing the reliability and interpretability of neural network outcomes. Our empirical validation, conducted across 10 real-world datasets, demonstrates the superiority of CF-T2NN over a wide array of state-of-the-art methods on various graph benchmarks. This work not only enhances the GNN framework with robust uncertainty quantification capabilities but also sets a new standard for reliability and precision in graph-structured data analysis.

cs.LG↗

arXiv2024

LoLDU: Low-Rank Adaptation via Lower-Diag-Upper Decomposition for Parameter-Efficient Fine-Tuning

The rapid growth of model scale has necessitated substantial computational resources for fine-tuning. Existing approach such as Low-Rank Adaptation (LoRA) has sought to address the problem of handling the large updated parameters in full fine-tuning. However, LoRA utilize random initialization and optimization of low-rank matrices to approximate updated weights, which can result in suboptimal convergence and an accuracy gap compared to full fine-tuning. To address these issues, we propose LoLDU, a Parameter-Efficient Fine-Tuning (PEFT) approach that significantly reduces trainable parameters by 2600 times compared to regular PEFT methods while maintaining comparable performance. LoLDU leverages Lower-Diag-Upper Decomposition (LDU) to initialize low-rank matrices for faster convergence and orthogonality. We focus on optimizing the diagonal matrix for scaling transformations. To the best of our knowledge, LoLDU has the fewest parameters among all PEFT approaches. We conducted extensive experiments across 4 instruction-following datasets, 6 natural language understanding (NLU) datasets, 8 image classification datasets, and image generation datasets with multiple model types (LLaMA2, RoBERTa, ViT, and Stable Diffusion), providing a comprehensive and detailed analysis. Our open-source code can be accessed at \href{https://github.com/SKDDJ/LoLDU}{https://github.com/SKDDJ/LoLDU}.

cs.CV↗

arXiv2024

An Eigengap Ratio Test for Determining the Number of Communities in Network Data

To characterize the community structure in network data, researchers have introduced various block-type models, including the stochastic block model, degree-corrected stochastic block model, mixed membership block model, degree-corrected mixed membership block model, and others. A critical step in applying these models effectively is determining the number of communities in the network. However, to our knowledge, existing methods for estimating the number of network communities often require model estimations or are unable to simultaneously account for network sparsity and a divergent number of communities. In this paper, we propose an eigengap-ratio based test that address these challenges. The test is straightforward to compute, requires no parameter tuning, and can be applied to a wide range of block models without the need to estimate network distribution parameters. Furthermore, it is effective for both dense and sparse networks with a divergent number of communities. We show that the proposed test statistic converges to a function of the type-I Tracy-Widom distributions under the null hypothesis, and that the test is asymptotically powerful under alternatives. Simulation studies on both dense and sparse networks demonstrate the efficacy of the proposed method. Three real-world examples are presented to illustrate the usefulness of the proposed test.

stat.ME↗

arXiv2024

SVFit: Parameter-Efficient Fine-Tuning of Large Pre-Trained Models Using Singular Values

Large pre-trained models (LPMs) have demonstrated exceptional performance in diverse natural language processing and computer vision tasks. However, fully fine-tuning these models poses substantial memory challenges, particularly in resource-constrained environments. Parameter-efficient fine-tuning (PEFT) methods, such as LoRA, mitigate this issue by adjusting only a small subset of parameters. Nevertheless, these methods typically employ random initialization for low-rank matrices, which can lead to inefficiencies in gradient descent and diminished generalizability due to suboptimal starting points. To address these limitations, we propose SVFit, a novel PEFT approach that leverages singular value decomposition (SVD) to initialize low-rank matrices using critical singular values as trainable parameters. Specifically, SVFit performs SVD on the pre-trained weight matrix to obtain the best rank-r approximation matrix, emphasizing the most critical singular values that capture over 99% of the matrix's information. These top-r singular values are then used as trainable parameters to scale the fundamental subspaces of the matrix, facilitating rapid domain adaptation. Extensive experiments across various pre-trained models in natural language understanding, text-to-image generation, and image classification tasks reveal that SVFit outperforms LoRA while requiring 16 times fewer trainable parameters.

cs.LG↗

arXiv2022

Inward and Outward Network Influence Analysis

Measuring heterogeneous influence across nodes in a network is critical in network analysis. This paper proposes an Inward and Outward Network Influence (IONI) model to assess nodal heterogeneity. Specifically, we allow for two types of influence parameters; one measures the magnitude of influence that each node exerts on others (outward influence), while we introduce a new parameter to quantify the receptivity of each node to being influenced by others (inward influence). Accordingly, these two types of influence measures naturally classify all nodes into four quadrants (high inward and high outward, low inward and high outward, low inward and low outward, high inward and low outward). To demonstrate our four-quadrant clustering method in practice, we apply the quasi-maximum likelihood approach to estimate the influence parameters, and we show the asymptotic properties of the resulting estimators. In addition, score tests are proposed to examine the homogeneity of the two types of influence parameters. To improve the accuracy of inferences about nodal influences, we introduce a Bayesian information criterion that selects the optimal influence model. The usefulness of the IONI model and the four-quadrant clustering method is illustrated via simulation studies and an empirical example involving customer segmentation.

stat.ME↗