arXiv subjects

Feng Li

Publications and source records attributed to Feng Li.

At least 91 records · Page 5Linked to original sources

arXiv2025

VQ-VA World: Towards High-Quality Visual Question-Visual Answering

This paper studies Visual Question-Visual Answering (VQ-VA): generating an image, rather than text, in response to a visual question -- an ability that has recently emerged in proprietary systems such as NanoBanana and GPT-Image. To also bring this capability to open-source models, we introduce VQ-VA World, a data-centric framework built around an agentic pipeline for large-scale, targeted data construction. Leveraging web-scale deployment, this pipeline crawls a massive amount of ~1.8M high-quality, interleaved image-text samples for model training. For evaluation, we further release IntelligentBench, a human-curated benchmark that systematically assesses VQ-VA along the aspects of world knowledge, design knowledge, and reasoning. Training with VQ-VA World data yields strong empirical gains: it helps LightFusion attain 53.06 on IntelligentBench, substantially surpassing the best prior open-source baselines (i.e., 7.78 from vanilla LightFusion; 1.94 from UniWorld-V1), and significantly narrowing the gap toward leading proprietary systems (e.g., 81.67 from NanoBanana; 82.64 from GPT-Image). By releasing the full suite of model weights, datasets, and pipelines, we hope to stimulate future research on VQ-VA.

cs.CV↗

arXiv2025

VALUE: Value-Aware Large Language Model for Query Rewriting via Weighted Trie in Sponsored Search

Query-to-bidword(i.e., bidding keyword) rewriting is fundamental to sponsored search, transforming noisy user queries into semantically relevant and commercially valuable keywords. Recent advances in large language models (LLMs) improve semantic relevance through generative retrieval frameworks, but they rarely encode the commercial value of keywords. As a result, rewrites are often semantically correct yet economically suboptimal, and a reinforcement learning from human feedback (RLHF) stage is usually added after supervised fine-tuning(SFT) to mitigate this deficiency. However, conventional preference alignment frequently overemphasize the ordering of bidword values and is susceptible to overfitting, which degrades rewrite quality. In addition, bidword value changes rapidly, while existing generative methods do not respond to these fluctuations. To address this shortcoming, we introduce VALUE(Value-Aware Large language model for qUery rewriting via wEighted trie), a framework that integrates value awareness directly into generation and enhances value alignment during training. VALUE employs the Weighted Trie, a novel variant of the classical trie that stores real-time value signals for each token. During decoding, the framework adjusts the LLM's token probabilities with these signals, constraining the search space and steering generation toward high-value rewrites. The alignment stage uses a fine-grained preference learning strategy that emphasizes stable, high-value differences and down-weights noisy or transient fluctuations, thereby improving robustness and reducing overfitting. Offline experiments show that VALUE significantly outperforms baselines in both semantic matching and value-centric metrics. VALUE has been deployed on our advertising system since October 2024 and served the Double Eleven promotions, the biggest shopping carnival in China.

cs.IR↗

arXiv2025

LightFusion: A Light-weighted, Double Fusion Framework for Unified Multimodal Understanding and Generation

Unified multimodal models have recently shown remarkable gains in both capability and versatility, yet most leading systems are still trained from scratch and require substantial computational resources. In this paper, we show that competitive performance can be obtained far more efficiently by strategically fusing publicly available models specialized for either generation or understanding. Our key design is to retain the original blocks while additionally interleaving multimodal self-attention blocks throughout the networks. This double fusion mechanism (1) effectively enables rich multi-modal fusion while largely preserving the original strengths of the base models, and (2) catalyzes synergistic fusion of high-level semantic representations from the understanding encoder with low-level spatial signals from the generation encoder. By training with only ~ 35B tokens, this approach achieves strong results across multiple benchmarks: 0.91 on GenEval for compositional text-to-image generation, 82.16 on DPG-Bench for complex text-to-image generation, 6.06 on GEditBench, and 3.77 on ImgEdit-Bench for image editing. By fully releasing the entire suite of code, model weights, and datasets, we hope to support future research on unified multimodal modeling.

cs.CV↗

arXiv2025

Multiwavelength Observations of the Apparently Non-repeating FRB 20250316A

The physical origin of fast radio bursts (FRBs) remains uncertain. Although multiwavelength observations have been widely conducted, only Galactic FRB~20200428D is associated with an X-ray burst from the magnetar SGR J1935+2154. Here, we present multiwavelength follow-up observations of the nearby bright FRB~20250316A, including the Five-hundred-meter Aperture Spherical radio Telescope (FAST), Einstein Probe (EP) X-ray mission, Chandra X-ray Observatory, Wide Field Survey Telescope (WFST) and Space Variable Object Monitor/Visible Telescope (SVOM/VT). The 13.08-hour FAST follow-up campaign without pulse detection requires an energy distribution flatter than those of well-known repeating FRBs, suggesting that this burst is likely a one-off event. A prompt EP follow-up and multi-epoch observational campaign totaling $>$ 100 ks led to the detection of an X-ray source within the angular resolution of its Follow-up X-ray Telescope (FXT, $10^{\prime\prime}$). A subsequent Chandra observation revealed this source to be offset by $7^{\prime\prime}$ from the FRB position, and established a 0.5-10 keV flux upper limit of $7.6\times 10^{-15}$ $\rm erg\,cm^{-2}\,s^{-1}$ at the FRB position, corresponding to $\sim 10^{39}$ $\rm erg\,s^{-1}$ at the 40 Mpc distance of the host galaxy NGC~4141. These results set one of the most stringent limits on X-ray emission from a non-repeating FRB, disfavoring ultra-luminous X-ray sources (ULXs) as counterparts of apparently one-off FRBs and offering critical insights into afterglow models. Our study suggests that an arcsecond localization of both the FRB and its potential X-ray counterpart is essential for exploring the X-ray counterpart of an FRB.

astro-ph.HE↗

arXiv2025

Tensor polarization and the dissipative damping of vector meson in QCD Medium

Unexpectedly large and puzzling spin alignment, and thus tensor polarization, of vector mesons has been observed in heavy-ion collisions. Given that tensor polarization represents a fluctuation of spin, we derive, for the first time, a fluctuation-dissipation relation for tensor polarization, where the polarization is proportional to first-order hydrodynamic gradients (e.g., the shear-stress tensor), with dissipative coefficients depending on the particle's damping properties, as characterized by its spectral function. Employing relativistic hydrodynamics at finite density, we find that dissipative contributions can generate substantial spin alignment. We provide illustrative examples (by tuning one coefficients $α_{\rm sh}$) that generate a beam energy, $p_T$, and centrality dependence of spin alignment resembling those observed in experiments, offering insights into these puzzling phenomena and demonstrating its potential as a ``spectrometer" for in-medium vector mesons.

nucl-th↗

arXiv2025

MCN-CL: Multimodal Cross-Attention Network and Contrastive Learning for Multimodal Emotion Recognition

Multimodal emotion recognition plays a key role in many domains, including mental health monitoring, educational interaction, and human-computer interaction. However, existing methods often face three major challenges: unbalanced category distribution, the complexity of dynamic facial action unit time modeling, and the difficulty of feature fusion due to modal heterogeneity. With the explosive growth of multimodal data in social media scenarios, the need for building an efficient cross-modal fusion framework for emotion recognition is becoming increasingly urgent. To this end, this paper proposes Multimodal Cross-Attention Network and Contrastive Learning (MCN-CL) for multimodal emotion recognition. It uses a triple query mechanism and hard negative mining strategy to remove feature redundancy while preserving important emotional cues, effectively addressing the issues of modal heterogeneity and category imbalance. Experiment results on the IEMOCAP and MELD datasets show that our proposed method outperforms state-of-the-art approaches, with Weighted F1 scores improving by 3.42% and 5.73%, respectively.

cs.CV↗

arXiv2025

FedeCouple: Fine-Grained Balancing of Global-Generalization and Local-Adaptability in Federated Learning

In privacy-preserving mobile network transmission scenarios with heterogeneous client data, personalized federated learning methods that decouple feature extractors and classifiers have demonstrated notable advantages in enhancing learning capability. However, many existing approaches primarily focus on feature space consistency and classification personalization during local training, often neglecting the local adaptability of the extractor and the global generalization of the classifier. This oversight results in insufficient coordination and weak coupling between the components, ultimately degrading the overall model performance. To address this challenge, we propose FedeCouple, a federated learning method that balances global generalization and local adaptability at a fine-grained level. Our approach jointly learns global and local feature representations while employing dynamic knowledge distillation to enhance the generalization of personalized classifiers. We further introduce anchors to refine the feature space; their strict locality and non-transmission inherently preserve privacy and reduce communication overhead. Furthermore, we provide a theoretical analysis proving that FedeCouple converges for nonconvex objectives, with iterates approaching a stationary point as the number of communication rounds increases. Extensive experiments conducted on five image-classification datasets demonstrate that FedeCouple consistently outperforms nine baseline methods in effectiveness, stability, scalability, and security. Notably, in experiments evaluating effectiveness, FedeCouple surpasses the best baseline by a significant margin of 4.3%.

cs.CV↗

arXiv2025

Prompt to Restore, Restore to Prompt: Cyclic Prompting for Universal Adverse Weather Removal

Universal adverse weather removal (UAWR) seeks to address various weather degradations within a unified framework. Recent methods are inspired by prompt learning using pre-trained vision-language models (e.g., CLIP), leveraging degradation-aware prompts to facilitate weather-free image restoration, yielding significant improvements. In this work, we propose CyclicPrompt, an innovative cyclic prompt approach designed to enhance the effectiveness, adaptability, and generalizability of UAWR. CyclicPrompt Comprises two key components: 1) a composite context prompt that integrates weather-related information and context-aware representations into the network to guide restoration. This prompt differs from previous methods by marrying learnable input-conditional vectors with weather-specific knowledge, thereby improving adaptability across various degradations. 2) The erase-and-paste mechanism, after the initial guided restoration, substitutes weather-specific knowledge with constrained restoration priors, inducing high-quality weather-free concepts into the composite prompt to further fine-tune the restoration process. Therefore, we can form a cyclic "Prompt-Restore-Prompt" pipeline that adeptly harnesses weather-specific knowledge, textual contexts, and reliable textures. Extensive experiments on synthetic and real-world datasets validate the superior performance of CyclicPrompt. The code is available at: https://github.com/RongxinL/CyclicPrompt.

cs.CV↗

arXiv2025

ROVER: Benchmarking Reciprocal Cross-Modal Reasoning for Omnimodal Generation

Unified multimodal models (UMMs) have emerged as a powerful paradigm for seamlessly unifying text and image understanding and generation. However, prevailing evaluations treat these abilities in isolation, such that tasks with multimodal inputs and outputs are scored primarily through unimodal reasoning, i.e., textual benchmarks emphasize language-based reasoning, while visual benchmarks emphasize reasoning outcomes manifested in the pixels. We introduce ROVER to address this pressing need to test reciprocal cross-modal reasoning, the use of one modality to guide, verify, or refine outputs in the other, an ability central to the vision of unified multimodal intelligence. ROVER is a human-annotated benchmark that explicitly targets reciprocal cross-modal reasoning, which contains 1312 tasks grounded in 1876 images, spanning two complementary settings. Verbally-augmented reasoning for visual generation evaluates whether models can use verbal prompts and reasoning chains to guide faithful image synthesis. Visually-augmented reasoning for verbal generation evaluates whether models can generate intermediate visualizations that strengthen their own reasoning processes for question answering. Experiments on 17 unified models reveal two key findings: (i) Cross-modal reasoning determines visual generation quality, with interleaved models significantly outperforming non-interleaved ones; notably, combining strong unimodal models fails to achieve comparable reasoning. (ii) Models show dissociation between physical and symbolic reasoning: they succeed at interpreting perceptual concepts literally but fail to construct visual abstractions for symbolic tasks, where faulty reasoning harms performance. These results highlight reciprocal cross-modal reasoning as a critical frontier for enabling true omnimodal generation.

cs.CV↗

arXiv2025

Towards Robust and Generalizable Continuous Space-Time Video Super-Resolution with Events

Continuous space-time video super-resolution (C-STVSR) has garnered increasing interest for its capability to reconstruct high-resolution and high-frame-rate videos at arbitrary spatial and temporal scales. However, prevailing methods often generalize poorly, producing unsatisfactory results when applied to out-of-distribution (OOD) scales. To overcome this limitation, we present EvEnhancer, a novel approach that marries the unique properties of high temporal resolution and high dynamic range encapsulated in event streams to achieve robust and generalizable C-STVSR. Our approach incorporates event-adapted synthesis that capitalizes on the spatiotemporal correlations between frames and events to capture long-term motion trajectories, enabling adaptive interpolation and fusion across space and time. This is then coupled with a local implicit video transformer that integrates local implicit video neural function with cross-scale spatiotemporal attention to learn continuous video representations and generate plausible videos at arbitrary resolutions and frame rates. We further develop EvEnhancerPlus, which builds a controllable switching mechanism that dynamically determines the reconstruction difficulty for each spatiotemporal pixel based on local event statistics. This allows the model to adaptively route reconstruction along the most suitable pathways at a fine-grained pixel level, substantially reducing computational overhead while maintaining excellent performance. Furthermore, we devise a cross-derivative training strategy that stabilizes the convergence of such a multi-pathway framework through staged cross-optimization. Extensive experiments demonstrate that our method achieves state-of-the-art performance on both synthetic and real-world datasets, while maintaining superior generalizability at OOD scales. The code is available at https://github.com/W-Shuoyan/EvEnhancerPlus.

eess.IV↗

arXiv2025

InstanceBEV: Unifying Instance and BEV Representation for 3D Panoptic Segmentation

BEV-based 3D perception has emerged as a focal point of research in end-to-end autonomous driving. However, existing BEV approaches encounter significant challenges due to the large feature space, complicating efficient modeling and hindering effective integration of global attention mechanisms. We propose a novel modeling strategy, called InstanceBEV, that synergistically combines the strengths of both map-centric approaches and object-centric approaches. Our method effectively extracts instance-level features within the BEV features, facilitating the implementation of global attention modeling in a highly compressed feature space, thereby addressing the efficiency challenges inherent in map-centric global modeling. Furthermore, our approach enables effective multi-task learning without introducing additional module. We validate the efficiency and accuracy of the proposed model through predicting occupancy, achieving 3D occupancy panoptic segmentation by combining instance information. Experimental results on the OCC3D-nuScenes dataset demonstrate that InstanceBEV, utilizing only 8 frames, achieves a RayPQ of 15.3 and a RayIoU of 38.2. This surpasses SparseOcc's RayPQ by 9.3% and RayIoU by 10.7%, showcasing the effectiveness of multi-task synergy.

cs.CV↗

arXiv2025

CLEAR: A Clinically-Grounded Tabular Framework for Radiology Report Evaluation

Existing metrics often lack the granularity and interpretability to capture nuanced clinical differences between candidate and ground-truth radiology reports, resulting in suboptimal evaluation. We introduce a Clinically-grounded tabular framework with Expert-curated labels and Attribute-level comparison for Radiology report evaluation (CLEAR). CLEAR not only examines whether a report can accurately identify the presence or absence of medical conditions, but also assesses whether it can precisely describe each positively identified condition across five key attributes: first occurrence, change, severity, descriptive location, and recommendation. Compared to prior works, CLEAR's multi-dimensional, attribute-level outputs enable a more comprehensive and clinically interpretable evaluation of report quality. Additionally, to measure the clinical alignment of CLEAR, we collaborate with five board-certified radiologists to develop CLEAR-Bench, a dataset of 100 chest X-ray reports from MIMIC-CXR, annotated across 6 curated attributes and 13 CheXpert conditions. Our experiments show that CLEAR achieves high accuracy in extracting clinical attributes and provides automated metrics that are strongly aligned with clinical judgment.

cs.CL↗

arXiv2025

On the Isomorphism Relation for Omnigenous Locally Finite Groups

The concept of an omnigenous locally finite group was introduced in [2] as a generalization of Hall's universal countable locally finite group. In this paper we show that the class of all countable omnigenous locally finite groups is Borel complete, hence it has the maximum Borel cardinality of isomorphism types among all countable structures. [2] M. Etedadialiabadi, S. Gao, F. Le Maître, J. Melleray, Dense locally finite subgroups of automorphism groups of ultraextensive spaces, Adv. Math. 391 (2021), 107966.

math.LO↗

arXiv2025

DOPA: Stealthy and Generalizable Backdoor Attacks from a Single Client under Challenging Federated Constraints

Federated Learning (FL) is increasingly adopted for privacy-preserving collaborative training, but its decentralized nature makes it particularly susceptible to backdoor attacks. Existing attack methods, however, often rely on idealized assumptions and fail to remain effective under real-world constraints, such as limited attacker control, non-IID data distributions, and the presence of diverse defense mechanisms. To address this gap, we propose DOPA (Divergent Optimization Path Attack), a novel framework that simulates heterogeneous local training dynamics and seeks consensus across divergent optimization trajectories to craft universally effective and stealthy backdoor triggers. By leveraging consistency signals across simulated paths to guide optimization, DOPA overcomes the challenge of heterogeneity-induced instability and achieves practical attack viability under stringent federated constraints. We validate DOPA on a comprehensive suite of 12 defense strategies, two model architectures (ResNet18/VGG16), two datasets (CIFAR-10/TinyImageNet), and both mild and extreme non-IID settings. Despite operating under a single-client, black-box, and sparsely participating threat model, DOPA consistently achieves high attack success, minimal accuracy degradation, low runtime, and long-term persistence. These results demonstrate a more practical attack paradigm, offering new perspectives for designing robust defense strategies in federated learning systems

cs.CR↗

arXiv2025

Solvability of Euler equations in the fractional Sobolev spaces in a bounded smooth domain

Euler equations are the basic system in fluid dynamics describing the motion of incompressible and inviscid ideal fluids. For a bounded smooth domain $Ω$ in $\mathbb{R}^n$. The well-posedness of Euler equations is well-known in Sobolev spaces $W^{k,p}(Ω)$ with the integer $k>\frac{n}{p}+1,\, 1<p<\infty$. In this article, we study the well-posedness of Euler equations in fractional Sobolev spaces on a bounded smooth domain. We first give a priori estimates of Euler equations in fractional Hilbert-Sobolev spaces by using the energy method. For the general case of fractional Sobolev spaces, we use the characteristic method together with elliptic estimates to give similar estimates. Finally, using the a priori estimate obtained we give solvability of Euler equations in fractional Sobolev spaces. Similar to the classical case, our result is global in time in the case of two dimensions and local in the three dimensions.

math.AP↗

arXiv2025

Hölder extension for fractional Laplacian

In this note, we characterize the sharp boundary condition such that the fractional harmonic extensions with Hölder regularity up to the boundary is globally Hölder continuous. The proofs are based on estimates of fractional harmonic measure decay and uniform fractional fatness of the complement of the domain.

math.AP↗

arXiv2025

Coulombic control of charge transfer in luminescent radicals with long-lived quartet states

Excitons in organic materials are emerging as an attractive platform for tunable quantum technologies. Structures with near-degenerate doublet and triplet excitations in linked trityl radical, acene and carbazole units can host quartet states. These high spin states can be coherently manipulated, and later decay radiatively via the radical doublet transition. However, this requires controlling the deexcitation pathways of all metastable states. Here we establish design rules for efficient quartet generation in luminescent radicals, using different connection arrangements of the molecular units. We discover that electronic coupling strength between these units dictates luminescence and quartet formation yields, particularly through the energetics of an acene-radical charge transfer state, which we tune Coulombically. This state acts as a source of non-radiative decay when acene-radical separation is small, but facilitates doublet-quartet spin interconversion when acene-radical separation is large. Using these rules we report a radical-carbazole-acene material with 55% luminescence yield, where 94% of emitting excitons originate from the quartet at microsecond times. This reveals the central role of molecular topology in luminescent quantum materials.

cond-mat.mtrl-sci↗

arXiv2025

Topological edge states and amplitude-dependent delocalization in quasiperiodic elliptically geared lattices

We present a class of mechanical lattices based on elliptical gears with quasiperiodic modulation and geometric nonlinearity, capable of exhibiting topologically protected modes and amplitude-driven transitions. Starting from a one-dimensional chain of modulated elliptical gears, we demonstrate the emergence of localized edge states arising from quasiperiodic variation in the gears' moments of inertia, analogous to the topological edge modes of the Aubry-Andre-Harper model. Under increasing excitation amplitude, the system undergoes a nonlinear transition, where edge localization breaks down and energy delocalizes into the bulk. By coupling multiple such chains with varying modulation phase, we construct a two-dimensional lattice in which the phase acts as a synthetic dimension. This structure supports topological wave propagation along the synthetic dimension. Nonlinearity again induces a breakdown of topological states, leading to complex, amplitude-dependent wave propagation. We further propose a numerical continuation approach to analyzing the periodic orbits and their linear stability, effectively discovering the boundary of the basin of bounded motion and detecting the occurrence of delocalization under certain excitation amplitudes. Our results reveal that elliptical geared systems offer a passive, amplitude-dependent platform for exploring topological phenomena and synthetic dimensionality in mechanical metamaterials.

physics.app-ph↗