arXiv subjects

Ziyu Liu

Publications and source records attributed to Ziyu Liu.

At least 55 records · Page 3Linked to original sources

arXiv2025

CoMA: Complementary Masking and Hierarchical Dynamic Multi-Window Self-Attention in a Unified Pre-training Framework

Masked Autoencoders (MAE) achieve self-supervised learning of image representations by randomly removing a portion of visual tokens and reconstructing the original image as a pretext task, thereby significantly enhancing pretraining efficiency and yielding excellent adaptability across downstream tasks. However, MAE and other MAE-style paradigms that adopt random masking generally require more pre-training epochs to maintain adaptability. Meanwhile, ViT in MAE suffers from inefficient parameter use due to fixed spatial resolution across layers. To overcome these limitations, we propose the Complementary Masked Autoencoders (CoMA), which employ a complementary masking strategy to ensure uniform sampling across all pixels, thereby improving effective learning of all features and enhancing the model's adaptability. Furthermore, we introduce DyViT, a hierarchical vision transformer that employs a Dynamic Multi-Window Self-Attention (DM-MSA), significantly reducing the parameters and FLOPs while improving fine-grained feature learning. Pre-trained on ImageNet-1K with CoMA, DyViT matches the downstream performance of MAE using only 12% of the pre-training epochs, demonstrating more effective learning. It also attains a 10% reduction in pre-training time per epoch, further underscoring its superior pre-training efficiency.

cs.CV↗

arXiv2025

Graph-Guided Concept Selection for Efficient Retrieval-Augmented Generation

Graph-based RAG constructs a knowledge graph (KG) from text chunks to enhance retrieval in Large Language Model (LLM)-based question answering. It is especially beneficial in domains such as biomedicine, law, and political science, where effective retrieval often involves multi-hop reasoning over proprietary documents. However, these methods demand numerous LLM calls to extract entities and relations from text chunks, incurring prohibitive costs at scale. Through a carefully designed ablation study, we observe that certain words (termed concepts) and their associated documents are more important. Based on this insight, we propose Graph-Guided Concept Selection (G2ConS). Its core comprises a chunk selection method and an LLM-independent concept graph. The former selects salient document chunks to reduce KG construction costs; the latter closes knowledge gaps introduced by chunk selection at zero cost. Evaluations on multiple real-world datasets show that G2ConS outperforms all baselines in construction cost, retrieval effectiveness, and answering quality.

cs.LG↗

arXiv2025

SPARK: Synergistic Policy And Reward Co-Evolving Framework

Recent Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) increasingly use Reinforcement Learning (RL) for post-pretraining, such as RL with Verifiable Rewards (RLVR) for objective tasks and RL from Human Feedback (RLHF) for subjective tasks. However, RLHF incurs high costs and potential reward-policy mismatch due to reliance on human preferences, while RLVR still wastes supervision by discarding rollouts and correctness signals after each update. To address these challenges, we introduce the Synergistic Policy And Reward Co-Evolving Framework (SPARK), an efficient, on-policy, and stable method that builds on RLVR. Instead of discarding rollouts and correctness data, SPARK recycles this valuable information to simultaneously train the model itself as a generative reward model. This auxiliary training uses a mix of objectives, such as pointwise reward score, pairwise comparison, and evaluation conditioned on further-reflection responses, to teach the model to evaluate and improve its own responses. Our process eliminates the need for a separate reward model and costly human preference data. SPARK creates a positive co-evolving feedback loop: improved reward accuracy yields better policy gradients, which in turn produce higher-quality rollouts that further refine the reward model. Our unified framework supports test-time scaling via self-reflection without external reward models and their associated costs. We show that SPARK achieves significant performance gains on multiple LLM and LVLM models and multiple reasoning, reward models, and general benchmarks. For example, SPARK-VL-7B achieves an average 9.7% gain on 7 reasoning benchmarks, 12.1% on 2 reward benchmarks, and 1.5% on 8 general benchmarks over the baselines, demonstrating robustness and broad generalization.

cs.CV↗

arXiv2025

Two-dimensional percolation model with long-range interaction

We perform large-scale simulations of the two-dimensional long-range bond percolation model with algebraically decaying percolation probabilities $\sim 1/r^{2+σ}$, using both conventional ensemble and event-based ensemble methods for system sizes up to $L=16384$. We accurately determine the critical points, the universal values of several dimensionless quantities, and the corresponding critical exponents. Our results provide compelling evidence that the system undergoes a crossover from short-range to long-range universality at $σ= 2$, in contradiction to Sak's criterion. Notably, we observe a pronounced jump in the universal values and critical exponents at $σ= 2$, a feature absent from previous studies.

cond-mat.stat-mech↗

arXiv2025

R&D of KLM Upgrade for Direct Measurement of Neutral Hadron Momentum via Time-of-Flight in Belle II

Accurate momentum determination of a neutral hadron, such as a KL meson or a neutron, remains a significant challenge in particle physics and nuclear physics experiments. The Belle II experiment presents an opportunity to address this challenge through an upgrade incorporating Time-of-Flight (TOF) capability for its large KL and Muon Detector (KLM). We investigate the feasibility of momentum determination via TOF measurement. To achieve high time resolution for the KLM upgrade, we conduct research and development of cost-effective plastic scintillators in collaboration with GaoNengKeDi Company, and technology utilizing silicon photomultipliers(SiPMs) arrays. A bulk attenuation length of 120 \pm 7 cm has been achieved with a 135 cm-long sample, along with a time resolution of 70 \pm 7 ps at its midpoint. A 50 cm-long scintillator demonstrates an exceptional time resolution of 47 \pm 2 ps. These results highlight the potential of the proposed technology for improving neutral hadron momentum measurements in an upgraded Belle II KLM detector.

physics.ins-det↗

arXiv2025

CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning

Autonomous agents for Graphical User Interfaces (GUIs) face significant challenges in specialized domains such as scientific computing, where both long-horizon planning and precise execution are required. Existing approaches suffer from a trade-off: generalist agents excel at planning but perform poorly in execution, while specialized agents demonstrate the opposite weakness. Recent compositional frameworks attempt to bridge this gap by combining a planner and an actor, but they are typically static and non-trainable, which prevents adaptation from experience. This is a critical limitation given the scarcity of high-quality data in scientific domains. To address these limitations, we introduce CODA, a novel and trainable compositional framework that integrates a generalist planner (Cerebrum) with a specialist executor (Cerebellum), trained via a dedicated two-stage pipeline. In the first stage, Specialization, we apply a decoupled GRPO approach to train an expert planner for each scientific application individually, bootstrapping from a small set of task trajectories. In the second stage, Generalization, we aggregate all successful trajectories from the specialized experts to build a consolidated dataset, which is then used for supervised fine-tuning of the final planner. This equips CODA with both robust execution and cross-domain generalization. Evaluated on four challenging applications from the ScienceBoard benchmark, CODA significantly outperforms baselines and establishes a new state of the art among open-source models.

cs.CV↗

arXiv2025

A Unified Contrastive-Generative Framework for Time Series Classification

Self-supervised learning (SSL) for multivariate time series mainly includes two paradigms: contrastive methods that excel at instance discrimination and generative approaches that model data distributions. While effective individually, their complementary potential remains unexplored. We propose a Contrastive Generative Time series framework (CoGenT), the first framework to unify these paradigms through joint contrastive-generative optimization. CoGenT addresses fundamental limitations of both approaches: it overcomes contrastive learning's sensitivity to high intra-class similarity in temporal data while reducing generative methods' dependence on large datasets. We evaluate CoGenT on six diverse time series datasets. The results show consistent improvements, with up to 59.2% and 14.27% F1 gains over standalone SimCLR and MAE, respectively. Our analysis reveals that the hybrid objective preserves discriminative power while acquiring generative robustness. These findings establish a foundation for hybrid SSL in temporal domains. We will release the code shortly.

cs.LG↗

arXiv2025

SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience

Repurposing large vision-language models (LVLMs) as computer use agents (CUAs) has led to substantial breakthroughs, primarily driven by human-labeled data. However, these models often struggle with novel and specialized software, particularly in scenarios lacking human annotations. To address this challenge, we propose SEAgent, an agentic self-evolving framework enabling CUAs to autonomously evolve through interactions with unfamiliar software. Specifically, SEAgent empowers computer-use agents to autonomously master novel software environments via experiential learning, where agents explore new software, learn through iterative trial-and-error, and progressively tackle auto-generated tasks organized from simple to complex. To achieve this goal, we design a World State Model for step-wise trajectory assessment, along with a Curriculum Generator that generates increasingly diverse and challenging tasks. The agent's policy is updated through experiential learning, comprised of adversarial imitation of failure actions and Group Relative Policy Optimization (GRPO) on successful ones. Furthermore, we introduce a specialist-to-generalist training strategy that integrates individual experiential insights from specialist agents, facilitating the development of a stronger generalist CUA capable of continuous autonomous evolution. This unified agent ultimately achieves performance surpassing ensembles of individual specialist agents on their specialized software. We validate the effectiveness of SEAgent across five novel software environments within OS-World. Our approach achieves a significant improvement of 23.2% in success rate, from 11.3% to 34.5%, over a competitive open-source CUA, i.e., UI-TARS.

cs.AI↗

arXiv2025

Dynamic Interfacial Quantum Dipoles in Charge Transfer Heterostructures

Hysteretic gate responses of two-dimensional material heterostructures serve as sensitive probes of the underlying electronic states and hold significant promise for the development of novel nanoelectronic devices. Here we identify a new mechanism of hysteretic behavior in graphene/$h$BN/$α$-$\mathrm{RuCl_3}$ charge transfer field effect devices. The hysteresis loop exhibits a sharp onset under low temperatures and evolves symmetrically relative to the charge transfer equilibrium. Unlike conventional flash memory devices, the charge transfer heterostructure features a transparent tunneling barrier and its hysteretic gate response is induced by the dynamic tuning of interfacial dipoles originating from quantum exchange interactions. The system acts effectively as a ferroelectric and gives rise to remarkable tunability of the hysteretic gate response under external electrical bias. Our work unveils a novel mechanism for engineering hysteretic behaviors via dynamic interfacial quantum dipoles.

cond-mat.mes-hall↗

arXiv2025

MambaVesselNet++: A Hybrid CNN-Mamba Architecture for Medical Image Segmentation

Medical image segmentation plays an important role in computer-aided diagnosis. Traditional convolution-based U-shape segmentation architectures are usually limited by the local receptive field. Existing vision transformers have been widely applied to diverse medical segmentation frameworks due to their superior capabilities of capturing global contexts. Despite the advantage, the real-world application of vision transformers is challenged by their non-linear self-attention mechanism, requiring huge computational costs. To address this issue, the selective state space model (SSM) Mamba has gained recognition for its adeptness in modeling long-range dependencies in sequential data, particularly noted for its efficient memory costs. In this paper, we propose MambaVesselNet++, a Hybrid CNN-Mamba framework for medical image segmentation. Our MambaVesselNet++ is comprised of a hybrid image encoder (Hi-Encoder) and a bifocal fusion decoder (BF-Decoder). In Hi-Encoder, we first devise the texture-aware layer to capture low-level semantic features by leveraging convolutions. Then, we utilize Mamba to effectively model long-range dependencies with linear complexity. The Bi-Decoder adopts skip connections to combine local and global information of the Hi-Encoder for the accurate generation of segmentation masks. Extensive experiments demonstrate that MambaVesselNet++ outperforms current convolution-based, transformer-based, and Mamba-based state-of-the-arts across diverse medical 2D, 3D, and instance segmentation tasks. The code is available at https://github.com/CC0117/MambaVesselNet.

cs.CV↗

arXiv2025

Alignment and Safety in Large Language Models: Safety Mechanisms, Training Paradigms, and Emerging Challenges

Due to the remarkable capabilities and growing impact of large language models (LLMs), they have been deeply integrated into many aspects of society. Thus, ensuring their alignment with human values and intentions has emerged as a critical challenge. This survey provides a comprehensive overview of practical alignment techniques, training protocols, and empirical findings in LLM alignment. We analyze the development of alignment methods across diverse paradigms, characterizing the fundamental trade-offs between core alignment objectives. Our analysis shows that while supervised fine-tuning enables basic instruction-following, preference-based methods offer more flexibility for aligning with nuanced human intent. We discuss state-of-the-art techniques, including Direct Preference Optimization (DPO), Constitutional AI, brain-inspired methods, and alignment uncertainty quantification (AUQ), highlighting their approaches to balancing quality and efficiency. We review existing evaluation frameworks and benchmarking datasets, emphasizing limitations such as reward misspecification, distributional robustness, and scalable oversight. We summarize strategies adopted by leading AI labs to illustrate the current state of practice. We conclude by outlining open problems in oversight, value pluralism, robustness, and continuous alignment. This survey aims to inform both researchers and practitioners navigating the evolving landscape of LLM alignment.

cs.AI↗

arXiv2025

CFFormer: Cross CNN-Transformer Channel Attention and Spatial Feature Fusion for Improved Segmentation of Heterogeneous Medical Images

Medical image segmentation plays an important role in computer-aided diagnosis. Existing methods mainly utilize spatial attention to highlight the region of interest. However, due to limitations of medical imaging devices, medical images exhibit significant heterogeneity, posing challenges for segmentation. Ultrasound images, for instance, often suffer from speckle noise, low resolution, and poor contrast between target tissues and background, which may lead to inaccurate boundary delineation. To address these challenges caused by heterogeneous image quality, we propose a hybrid CNN-Transformer model,called CFFormer, which leverages effective channel feature extraction to enhance the model' s ability to accurately identify tissue regions by capturing rich contextual information. The proposed architecture contains two key components: the Cross Feature Channel Attention (CFCA) module and the X-Spatial Feature Fusion (XFF) module. The model incorporates dual encoders, with the CNN encoder focusing on capturing local features and the Transformer encoder modeling global features. The CFCA module filters and facilitates interactions between the channel features from the two encoders, while the XFF module effectively reduces the significant semantic information differences in spatial features, enabling a smooth and cohesive spatial feature fusion. We evaluate our model across eight datasets covering five modalities to test its generalization capability. Experimental results demonstrate that our model outperforms current state-of-the-art methods and maintains accurate tissue region segmentation across heterogeneous medical image datasets. The code is available at https://github.com/JiaxuanFelix/CFFormer.

cs.CV↗

arXiv2025

InternLM-XComposer2.5-Reward: A Simple Yet Effective Multi-Modal Reward Model

Despite the promising performance of Large Vision Language Models (LVLMs) in visual understanding, they occasionally generate incorrect outputs. While reward models (RMs) with reinforcement learning or test-time scaling offer the potential for improving generation quality, a critical gap remains: publicly available multi-modal RMs for LVLMs are scarce, and the implementation details of proprietary models are often unclear. We bridge this gap with InternLM-XComposer2.5-Reward (IXC-2.5-Reward), a simple yet effective multi-modal reward model that aligns LVLMs with human preferences. To ensure the robustness and versatility of IXC-2.5-Reward, we set up a high-quality multi-modal preference corpus spanning text, image, and video inputs across diverse domains, such as instruction following, general understanding, text-rich documents, mathematical reasoning, and video understanding. IXC-2.5-Reward achieves excellent results on the latest multi-modal reward model benchmark and shows competitive performance on text-only reward model benchmarks. We further demonstrate three key applications of IXC-2.5-Reward: (1) Providing a supervisory signal for RL training. We integrate IXC-2.5-Reward with Proximal Policy Optimization (PPO) yields IXC-2.5-Chat, which shows consistent improvements in instruction following and multi-modal open-ended dialogue; (2) Selecting the best response from candidate responses for test-time scaling; and (3) Filtering outlier or noisy samples from existing image and video instruction tuning training data. To ensure reproducibility and facilitate further research, we have open-sourced all model weights and training recipes at https://github.com/InternLM/InternLM-XComposer/tree/main/InternLM-XComposer-2.5-Reward

cs.CV↗

arXiv2025

Visual Agentic Reinforcement Fine-Tuning

A key trend in Large Reasoning Models (e.g., OpenAI's o3) is the native agentic ability to use external tools such as web browsers for searching and writing/executing code for image manipulation to think with images. In the open-source research community, while significant progress has been made in language-only agentic abilities such as function calling and tool integration, the development of multi-modal agentic capabilities that involve truly thinking with images, and their corresponding benchmarks, are still less explored. This work highlights the effectiveness of Visual Agentic Reinforcement Fine-Tuning (Visual-ARFT) for enabling flexible and adaptive reasoning abilities for Large Vision-Language Models (LVLMs). With Visual-ARFT, open-source LVLMs gain the ability to browse websites for real-time information updates and write code to manipulate and analyze input images through cropping, rotation, and other image processing techniques. We also present a Multi-modal Agentic Tool Bench (MAT) with two settings (MAT-Search and MAT-Coding) designed to evaluate LVLMs' agentic search and coding abilities. Our experimental results demonstrate that Visual-ARFT outperforms its baseline by +18.6% F1 / +13.0% EM on MAT-Coding and +10.3% F1 / +8.7% EM on MAT-Search, ultimately surpassing GPT-4o. Visual-ARFT also achieves +29.3 F1% / +25.9% EM gains on existing multi-hop QA benchmarks such as 2Wiki and HotpotQA, demonstrating strong generalization capabilities. Our findings suggest that Visual-ARFT offers a promising path toward building robust and generalizable multimodal agents.

cs.CV↗

arXiv2025

Dance of Fireworks: An Interactive Broadcast Gymnastics Training System Based on Pose Estimation

This study introduces Dance of Fireworks, an interactive system designed to combat sedentary health risks by enhancing engagement in radio calisthenics. Leveraging mobile device cameras and lightweight pose estimation (PoseNet/TensorFlow Lite), the system extracts body keypoints, computes joint angles, and compares them with standardized motions to deliver real-time corrective feedback. To incentivize participation, it dynamically maps users' movements (such as joint angles and velocity) to customizable fireworks animations, rewarding improved accuracy with richer visual effects. Experiments involving 136 participants demonstrated a significant reduction in average joint angle errors from 21.3 degrees to 9.8 degrees (p < 0.01) over four sessions, with 93.4 percent of users affirming its exercise-promoting efficacy and 85.4 percent praising its entertainment value. The system operates without predefined motion templates or specialised hardware, enabling seamless integration into office environments. Future enhancements will focus on improving pose recognition accuracy, reducing latency, and adding features such as multiplayer interaction and music synchronisation. This work presents a cost-effective, engaging solution to promote physical activity in sedentary populations.

cs.CV↗

arXiv2025

Cost-based Selection of Provenance Sketches for Data Skipping

Provenance sketches, light-weight indexes that record what data is needed (is relevant) for answering a query, can significantly improve performance of important classes of queries (e.g., HAVING and top-k queries). Given a horizontal partition of a table, a provenance sketch for a query Q records which fragments contain provenance. Once a provenance sketch has been captured for a query, it can be used to speed-up subsequent queries by skipping data that does not belong to a sketch. The size and, thus, also the effectiveness of a provenance sketch is often quite sensitive to the choice of attribute(s) we are partitioning on. In this work, we develop sample-based estimation techniques for the size of provenance sketches akin to a specialized form of approximate query processing. This technique enables the online selection of provenance sketches by estimating the size of sketches for a set of candidate attributes and then creating the sketch that is estimated to yield the largest benefit. We demonstrate experimentally that our estimation is accurate enough to select optimal or near optimal provenance sketches in most cases which in turn leads to a runtime improvement of up to %60 compared to other strategies for selecting provenance sketches.

cs.DB↗

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

AD-GPT: Large Language Models in Alzheimer's Disease

Large language models (LLMs) have emerged as powerful tools for medical information retrieval, yet their accuracy and depth remain limited in specialized domains such as Alzheimer's disease (AD), a growing global health challenge. To address this gap, we introduce AD-GPT, a domain-specific generative pre-trained transformer designed to enhance the retrieval and analysis of AD-related genetic and neurobiological information. AD-GPT integrates diverse biomedical data sources, including potential AD-associated genes, molecular genetic information, and key gene variants linked to brain regions. We develop a stacked LLM architecture combining Llama3 and BERT, optimized for four critical tasks in AD research: (1) genetic information retrieval, (2) gene-brain region relationship assessment, (3) gene-AD relationship analysis, and (4) brain region-AD relationship mapping. Comparative evaluations against state-of-the-art LLMs demonstrate AD-GPT's superior precision and reliability across these tasks, underscoring its potential as a robust and specialized AI tool for advancing AD research and biomarker discovery.

cs.CL↗