arXiv subjects

Feng Zhang

Publications and source records attributed to Feng Zhang.

At least 91 records · Page 5Linked to original sources

arXiv2025

Ab initio superionic-liquid phase diagram of Fe1-xOx under Earth's inner core conditions

The superionic state is a phase of matter in which liquid-like ionic mobility coexists with a solid crystalline lattice. Recently identified in Earth's inner core (IC), this state has attracted considerable attention for its unique kinetic behavior and geophysical implications. However, the ab initio phase diagram describing the equilibrium between the superionic phase and the liquid solution under core conditions remains largely unexplored. Here, we present a thermodynamic approach to compute the Gibbs free energy and construct the ab initio superionic-liquid phase diagram for the Fe1-xOx system under IC conditions. We find that oxygen forms superionic states in both hcp and bcc Fe phases, with a pronounced influence on cooperative diffusion of iron in the bcc lattice. The stability fields of these superionic phases are sensitive to oxygen stoichiometry. The presence of superionic states leads to a higher oxygen concentration in the IC than previously estimated. Our work establishes a framework for investigating superionic-liquid equilibria under extreme conditions.

physics.geo-ph↗

arXiv2025

Thinking with Programming Vision: Towards a Unified View for Thinking with Images

Multimodal large language models (MLLMs) that think with images can interactively use tools to reason about visual inputs, but current approaches often rely on a narrow set of tools with limited real-world necessity and scalability. In this work, we first reveal a critical and previously overlooked weakness: even state-of-the-art MLLMs are surprisingly brittle, showing significant performance degradation on images with simple orientation changes or natural corruptions, underscoring the need for more robust tool-based reasoning. To address this, we propose CodeVision, a flexible and scalable code-as-tool framework where the model generates code as a universal interface to invoke any image operation, moving beyond fixed tool registries. We train our model using a two-stage methodology, beginning with Supervised Fine-Tuning (SFT) on a high-quality dataset curated for complex, multi-turn tool composition and error recovery, followed by Reinforcement Learning (RL) with a novel and dense process reward function to encourage strategic and efficient tool use. To facilitate this research, we construct new SFT and RL datasets and introduce a challenging new benchmark suite designed to rigorously evaluate robustness to orientation changes and multi-tool reasoning. Experiments on Qwen2.5-VL and Qwen3-VL series show that our approach significantly improves model performance and fosters emergent capabilities such as flexible tool composition, efficient chained execution, and robust error recovery from runtime feedback. Code is available at https://github.com/ByteDance-BandAI/CodeVision.

cs.CV↗

arXiv2025

Cross-Layer Vision Smoothing: Enhancing Visual Understanding via Sustained Focus on Key Objects in Large Vision-Language Models

Large Vision-Language Models (LVLMs) can accurately locate key objects in images, yet their attention to these objects tends to be very brief. Motivated by the hypothesis that sustained focus on key objects can improve LVLMs' visual capabilities, we propose Cross-Layer Vision Smoothing (CLVS). The core idea of CLVS is to incorporate a vision memory that smooths the attention distribution across layers. Specifically, we initialize this vision memory with position-unbiased visual attention in the first layer. In subsequent layers, the model's visual attention jointly considers the vision memory from previous layers, while the memory is updated iteratively, thereby maintaining smooth attention on key objects. Given that visual understanding primarily occurs in the early and middle layers of the model, we use uncertainty as an indicator of completed visual understanding and terminate the smoothing process accordingly. Experiments on four benchmarks across three LVLMs confirm the effectiveness and generalizability of our method. CLVS achieves state-of-the-art overall performance across a variety of visual understanding tasks and attains comparable results to the leading approaches on image captioning benchmarks.

cs.CV↗

arXiv2025

VoiceAgentEval: A Dual-Dimensional Benchmark for Expert-Level Intelligent Voice-Agent Evaluation of Xbench's Professional-Aligned Series

We propose OutboundEval, a comprehensive benchmark for evaluating large language models (LLMs) in expert-level intelligent outbound calling scenarios. Unlike existing methods that suffer from three key limitations - insufficient dataset diversity and category coverage, unrealistic user simulation, and inaccurate evaluation metrics - OutboundEval addresses these issues through a structured framework. First, we design a benchmark spanning six major business domains and 30 representative sub-scenarios, each with scenario-specific process decomposition, weighted scoring, and domain-adaptive metrics. Second, we develop a large-model-driven User Simulator that generates diverse, persona-rich virtual users with realistic behaviors, emotional variability, and communication styles, providing a controlled yet authentic testing environment. Third, we introduce a dynamic evaluation method that adapts to task variations, integrating automated and human-in-the-loop assessment to measure task execution accuracy, professional knowledge application, adaptability, and user experience quality. Experiments on 12 state-of-the-art LLMs reveal distinct trade-offs between expert-level task completion and interaction fluency, offering practical insights for building reliable, human-like outbound AI systems. OutboundEval establishes a practical, extensible, and domain-oriented standard for benchmarking LLMs in professional applications.

cs.AI↗

arXiv2025

LGM-Pose: A Lightweight Global Modeling Network for Real-time Human Pose Estimation

Most of the current top-down multi-person pose estimation lightweight methods are based on multi-branch parallel pure CNN network architecture, which often struggle to capture the global context required for detecting semantically complex keypoints and are hindered by high latency due to their intricate and redundant structures. In this article, an approximate single-branch lightweight global modeling network (LGM-Pose) is proposed to address these challenges. In the network, a lightweight MobileViM Block is designed with a proposed Lightweight Attentional Representation Module (LARM), which integrates information within and between patches using the Non-Parametric Transformation Operation(NPT-Op) to extract global information. Additionally, a novel Shuffle-Integrated Fusion Module (SFusion) is introduced to effectively integrate multi-scale information, mitigating performance degradation often observed in single-branch structures. Experimental evaluations on the COCO and MPII datasets demonstrate that our approach not only reduces the number of parameters compared to existing mainstream lightweight methods but also achieves superior performance and faster processing speeds.

cs.CV↗

arXiv2025

High-resolution ensemble retrieval of cloud properties for all-day based on geostationary satellite

Clouds play a critical role in Earth's hydrological and energy cycles, and accurately representing their properties is essential for effective numerical modeling and weather forecasting. Machine learning methods have been widely used for cloud property retrieval; however, most existing techniques are deterministic and do not incorporate uncertainty quantification. Generative machine learning has made significant advances in various domains, including natural language processing, image generation, and notably weather forecasting, where it has enabled ensemble predictions and the quantification of forecast uncertainty. This ability to quantify uncertainty offers valuable opportunities for cloud remote sensing. In this study, we propose a novel cloud property retrieval method, CloudDiff, based on a generative diffusion model. By leveraging thermal infrared observations from the Himawari-8 Advanced Himawari Imager (AHI), CloudDiff generates high spatiotemporal resolution cloud properties for both daytime and nighttime conditions, increasing the resolution of Himawari-8/AHI cloud retrievals from 2 km to 1 km. Unlike deterministic retrieval methods, CloudDiff generates multiple samples from the underlying probability distribution, allowing for a diverse range of plausible retrievals and taking steps towards providing uncertainty assessment. Additionally, CloudDiff produces sharper samples and better captures fine local features, enhancing the precision of cloud property retrieval. By averaging over the ensemble of generated samples, we demonstrate that both the accuracy and reliability of the retrievals are significantly improved. These high-resolution cloud properties have been successfully applied to analyze extreme weather events, such as typhoons, providing potentially valuable insights into atmospheric processes.

physics.ao-ph↗

arXiv2025

FlashMP: Fast Discrete Transform-Based Solver for Preconditioning Maxwell's Equations on GPUs

Efficiently solving large-scale linear systems is a critical challenge in electromagnetic simulations, particularly when using the Crank-Nicolson Finite-Difference Time-Domain (CN-FDTD) method. Existing iterative solvers are commonly employed to handle the resulting sparse systems but suffer from slow convergence due to the ill-conditioned nature of the double-curl operator. Approximate preconditioners, like Successive Over-Relaxation (SOR) and Incomplete LU decomposition (ILU), provide insufficient convergence, while direct solvers are impractical due to excessive memory requirements. To address this, we propose FlashMP, a novel preconditioning system that designs a subdomain exact solver based on discrete transforms. FlashMP provides an efficient GPU implementation that achieves multi-GPU scalability through domain decomposition. Evaluations on AMD MI60 GPU clusters (up to 1000 GPUs) show that FlashMP reduces iteration counts by up to 16x and achieves speedups of 2.5x to 4.9x compared to baseline implementations in state-of-the-art libraries such as Hypre. Weak scalability tests show parallel efficiencies up to 84.1%.

cs.DC↗

arXiv2025

Hyperspectral Anomaly Detection Fused Unified Nonconvex Tensor Ring Factors Regularization

In recent years, tensor decomposition-based approaches for hyperspectral anomaly detection (HAD) have gained significant attention in the field of remote sensing. However, existing methods often fail to fully leverage both the global correlations and local smoothness of the background components in hyperspectral images (HSIs), which exist in both the spectral and spatial domains. This limitation results in suboptimal detection performance. To mitigate this critical issue, we put forward a novel HAD method named HAD-EUNTRFR, which incorporates an enhanced unified nonconvex tensor ring (TR) factors regularization. In the HAD-EUNTRFR framework, the raw HSIs are first decomposed into background and anomaly components. The TR decomposition is then employed to capture the spatial-spectral correlations within the background component. Additionally, we introduce a unified and efficient nonconvex regularizer, induced by tensor singular value decomposition (TSVD), to simultaneously encode the low-rankness and sparsity of the 3-D gradient TR factors into a unique concise form. The above characterization scheme enables the interpretable gradient TR factors to inherit the low-rankness and smoothness of the original background. To further enhance anomaly detection, we design a generalized nonconvex regularization term to exploit the group sparsity of the anomaly component. To solve the resulting doubly nonconvex model, we develop a highly efficient optimization algorithm based on the alternating direction method of multipliers (ADMM) framework. Experimental results on several benchmark datasets demonstrate that our proposed method outperforms existing state-of-the-art (SOTA) approaches in terms of detection accuracy.

cs.CV↗

arXiv2025

Towards Flash Thinking via Decoupled Advantage Policy Optimization

Recent Large Reasoning Models (LRMs) have achieved remarkable performance in solving complex problems via supervised fine-tuning (SFT) and reinforcement learning (RL). Although existing RL algorithms significantly enhance model accuracy, they still suffer from excessively lengthy responses and overthinking issues, resulting in increased inference latency and computational consumption, especially for simple tasks that require minimal reasoning. To address this, we propose a novel RL framework, DEPO, to reduce inefficient reasoning for models. Our method mainly consists of three core components: (1) an innovative advantage decoupled algorithm to guide model reduction of inefficient tokens; (2) a difficulty-aware length penalty to lower the overall length of model responses; (3) an advantage clipping method to prevent bias in policy optimization. In our experiments, applied to DeepSeek-Distill-Qwen-7B and DeepSeek-Distill-Qwen-1.5B as base models, DEPO achieves a significant reduction in sequence length by 39% and reduces excessive reasoning paths in inefficient tokens, while outperforming the base model in overall accuracy.

cs.AI↗

arXiv2025

High-resolution Photo Enhancement in Real-time: A Laplacian Pyramid Network

Photo enhancement plays a crucial role in augmenting the visual aesthetics of a photograph. In recent years, photo enhancement methods have either focused on enhancement performance, producing powerful models that cannot be deployed on edge devices, or prioritized computational efficiency, resulting in inadequate performance for real-world applications. To this end, this paper introduces a pyramid network called LLF-LUT++, which integrates global and local operators through closed-form Laplacian pyramid decomposition and reconstruction. This approach enables fast processing of high-resolution images while also achieving excellent performance. Specifically, we utilize an image-adaptive 3D LUT that capitalizes on the global tonal characteristics of downsampled images, while incorporating two distinct weight fusion strategies to achieve coarse global image enhancement. To implement this strategy, we designed a spatial-frequency transformer weight predictor that effectively extracts the desired distinct weights by leveraging frequency features. Additionally, we apply local Laplacian filters to adaptively refine edge details in high-frequency components. After meticulously redesigning the network structure and transformer model, LLF-LUT++ not only achieves a 2.64 dB improvement in PSNR on the HDR+ dataset, but also further reduces runtime, with 4K resolution images processed in just 13 ms on a single GPU. Extensive experimental results on two benchmark datasets further show that the proposed approach performs favorably compared to state-of-the-art methods. The source code will be made publicly available at https://github.com/fengzhang427/LLF-LUT.

cs.CV↗

arXiv2025

Ab initio calculation of atomic solid hydrogen phases based on Gutzwiller many-body wave functions

We apply two ab initio many-body methods based on Gutzwiller wave functions, i.e., correlation matrix renormalization theory (CMRT) and Gutzwiller conjugate gradient minimization (GCGM), to the study of crystalline phases of atomic hydrogen. Both methods avoid empirical Hubbard U parameters and are free from double-counting issues. CMRT employs a Gutzwiller-type approximation that enables efficient calculations, while GCGM goes beyond this approximation to achieve higher accuracy at higher computational cost. By benchmarking against available quantum Monte Carlo results, we demonstrate that while both methods are more accurate than the widely used density-functional theory (DFT), GCGM systematically captures additional correlation energy missing in CMRT, leading to significantly improved total energy predictions. We also show that by including the correlation energy $E_c$ from LDA in the CMRT calculation, CMRT+$E_c$ produces energy in better agreement with the QMC results in these hydrogen lattice systems.

cond-mat.str-el↗

arXiv2025

Mitigating Hallucination in Large Vision-Language Models through Aligning Attention Distribution to Information Flow

Due to the unidirectional masking mechanism, Decoder-Only models propagate information from left to right. LVLMs (Large Vision-Language Models) follow the same architecture, with visual information gradually integrated into semantic representations during forward propagation. Through systematic analysis, we observe that the majority of the visual information is absorbed into the semantic representations. However, the model's attention distribution does not exhibit sufficient emphasis on semantic representations. This misalignment between the attention distribution and the actual information flow undermines the model's visual understanding ability and contributes to hallucinations. To address this issue, we enhance the model's visual understanding by leveraging the core information embedded in semantic representations. Specifically, we identify attention heads that focus on core semantic representations based on their attention distributions. Then, through a two-stage optimization paradigm, we propagate the advantages of these attention heads across the entire model, aligning the attention distribution with the actual information flow. We evaluate our method on three image captioning benchmarks using five different LVLMs, demonstrating its effectiveness in significantly reducing hallucinations. Further experiments reveal a trade-off between reduced hallucinations and richer details. Notably, our method allows for manual adjustment of the model's conservativeness, enabling flexible control to meet diverse real-world requirements.

cs.CV↗

arXiv2025

FastCuRL: Curriculum Reinforcement Learning with Stage-wise Context Scaling for Efficient Training R1-like Reasoning Models

Improving training efficiency continues to be one of the primary challenges in large-scale Reinforcement Learning (RL). In this paper, we investigate how context length and the complexity of training data influence the RL scaling training process of R1-distilled reasoning models, e.g., DeepSeek-R1-Distill-Qwen-1.5B. Our experimental results reveal that: (1) simply controlling the context length and curating the training data based on the input prompt length can effectively improve the training efficiency of RL scaling, achieving better performance with more concise CoT; (2) properly scaling the context length helps mitigate entropy collapse; and (3) carefully choosing the context length facilitates achieving efficient LLM training and reasoning. Inspired by these insights, we propose FastCuRL, a curriculum RL framework with stage-wise context scaling to achieve efficient LLM training and reasoning. Extensive experimental results demonstrate that FastCuRL-1.5B-V3 significantly outperforms state-of-the-art reasoning models on five competition-level benchmarks and achieves 49.6% accuracy on AIME 2024. Furthermore, FastCuRL-1.5B-Preview surpasses DeepScaleR-1.5B-Preview on five benchmarks while only using a single node with 8 GPUs and a total of 50% of training steps.

cs.CL↗

arXiv2025

Chain-of-Thought Re-ranking for Image Retrieval Tasks

Image retrieval remains a fundamental yet challenging problem in computer vision. While recent advances in Multimodal Large Language Models (MLLMs) have demonstrated strong reasoning capabilities, existing methods typically employ them only for evaluation, without involving them directly in the ranking process. As a result, their rich multimodal reasoning abilities remain underutilized, leading to suboptimal performance. In this paper, we propose a novel Chain-of-Thought Re-Ranking (CoTRR) method to address this issue. Specifically, we design a listwise ranking prompt that enables MLLM to directly participate in re-ranking candidate images. This ranking process is grounded in an image evaluation prompt, which assesses how well each candidate aligns with users query. By allowing MLLM to perform listwise reasoning, our method supports global comparison, consistent reasoning, and interpretable decision-making - all of which are essential for accurate image retrieval. To enable structured and fine-grained analysis, we further introduce a query deconstruction prompt, which breaks down the original query into multiple semantic components. Extensive experiments on five datasets demonstrate the effectiveness of our CoTRR method, which achieves state-of-the-art performance across three image retrieval tasks, including text-to-image retrieval (TIR), composed image retrieval (CIR) and chat-based image retrieval (Chat-IR). Our code is available at https://github.com/freshfish15/CoTRR .

cs.CV↗

arXiv2025

LLM-I: LLMs are Naturally Interleaved Multimodal Creators

We propose LLM-Interleaved (LLM-I), a flexible and dynamic framework that reframes interleaved image-text generation as a tool-use problem. LLM-I is designed to overcome the "one-tool" bottleneck of current unified models, which are limited to synthetic imagery and struggle with tasks requiring factual grounding or programmatic precision. Our framework empowers a central LLM or MLLM agent to intelligently orchestrate a diverse toolkit of specialized visual tools, including online image search, diffusion-based generation, code execution, and image editing. The agent is trained to select and apply these tools proficiently via a Reinforcement Learning (RL) framework that features a hybrid reward system combining rule-based logic with judgments from LLM and MLLM evaluators. Trained on a diverse new dataset using four different model backbones, LLM-I demonstrates state-of-the-art performance, outperforming existing methods by a large margin across four benchmarks. We also introduce a novel test-time scaling strategy that provides further performance gains. Project Page: https://github.com/ByteDance-BandAI/LLM-I.

cs.LG↗

arXiv2025

Cross-Image Contrastive Decoding: Precise, Lossless Suppression of Language Priors in Large Vision-Language Models

Over-reliance on language priors is a major cause of hallucinations in Large Vision-Language Models (LVLMs), often leading to outputs that are linguistically plausible but visually inconsistent. Recent studies have explored contrastive decoding as a training-free solution. However, these methods typically construct contrastive visual inputs by perturbing the original image, resulting in distorted contrastive distributions, incomplete contrastive signals, and excessive suppression of language priors. Motivated by the observation that language priors tend to remain consistent across different images, we propose Cross-Image Contrastive Decoding (CICD), a simple yet effective training-free method that uses unrelated images as contrastive visual inputs. To address the issue of over-suppressing language priors, which can negatively affect the quality of generated responses, we further introduce a dynamic selection mechanism based on the cross-image differences in model behavior. By selectively suppressing language priors, our method reduces hallucinations without compromising the model's performance. Extensive experiments across multiple benchmarks and LVLMs confirm the effectiveness and generalizability of CICD, particularly in image captioning, where language priors are especially dominant.

cs.CV↗

arXiv2025

CAMB: A comprehensive industrial LLM benchmark on civil aviation maintenance

Civil aviation maintenance is a domain characterized by stringent industry standards. Within this field, maintenance procedures and troubleshooting represent critical, knowledge-intensive tasks that require sophisticated reasoning. To address the lack of specialized evaluation tools for large language models (LLMs) in this vertical, we propose and develop an industrial-grade benchmark specifically designed for civil aviation maintenance. This benchmark serves a dual purpose: It provides a standardized tool to measure LLM capabilities within civil aviation maintenance, identifying specific gaps in domain knowledge and complex reasoning. By pinpointing these deficiencies, the benchmark establishes a foundation for targeted improvement efforts (e.g., domain-specific fine-tuning, RAG optimization, or specialized prompt engineering), ultimately facilitating progress toward more intelligent solutions within civil aviation maintenance. Our work addresses a significant gap in the current LLM evaluation, which primarily focuses on mathematical and coding reasoning tasks. In addition, given that Retrieval-Augmented Generation (RAG) systems are currently the dominant solutions in practical applications , we leverage this benchmark to evaluate existing well-known vector embedding models and LLMs for civil aviation maintenance scenarios. Through experimental exploration and analysis, we demonstrate the effectiveness of our benchmark in assessing model performance within this domain, and we open-source this evaluation benchmark and code to foster further research and development:https://github.com/CamBenchmark/cambenchmark

cs.CL↗

arXiv2025

VQualA 2025 Challenge on Face Image Quality Assessment: Methods and Results

Face images play a crucial role in numerous applications; however, real-world conditions frequently introduce degradations such as noise, blur, and compression artifacts, affecting overall image quality and hindering subsequent tasks. To address this challenge, we organized the VQualA 2025 Challenge on Face Image Quality Assessment (FIQA) as part of the ICCV 2025 Workshops. Participants created lightweight and efficient models (limited to 0.5 GFLOPs and 5 million parameters) for the prediction of Mean Opinion Scores (MOS) on face images with arbitrary resolutions and realistic degradations. Submissions underwent comprehensive evaluations through correlation metrics on a dataset of in-the-wild face images. This challenge attracted 127 participants, with 1519 final submissions. This report summarizes the methodologies and findings for advancing the development of practical FIQA approaches.

cs.CV↗