arXiv subjects

Guangyao Li

Publications and source records attributed to Guangyao Li.

At least 19 recordsLinked to original sources

arXiv2026

AVI-Bench: Toward Human-like Audio-Visual Intelligence of Omni-MLLMs

Recent advances in Omni-Multimodal Large Language Models (Omni-MLLMs) have enabled strong integration of vision, audio, and language. However, their audio-visual intelligence (AVI) remains insufficiently evaluated due to the lack of systematic and comprehensive benchmarks. We introduce AVI-Bench, a cognitively inspired benchmark that evaluates Omni-MLLMs across three stages, perception, understanding, and reasoning, through cross-modal tasks requiring joint audio-visual interpretation. This design enables fine-grained diagnosis of model capabilities and failure modes. To further assess robustness beyond familiar domains, we propose AVI-Bench-PriSe, an extension that probes models' primitive audio-visual sensation using unfamiliar, low-semantic stimuli, testing generalization beyond common training distributions. Extensive experiments on both open-source and closed-source models reveal substantial limitations in current Omni-MLLMs. Based on these findings, we present a four-level AVI taxonomy. Overall, AVI-Bench provides a principled evaluation framework to guide the development of more robust and generalizable AVI. Project website: https://fudancvl.github.io/AVI-Bench/

cs.CV↗

arXiv2026

LEGO-MOF: Equivariant Latent Manipulation for Editable, Generative, and Optimizable MOF Design

Metal-organic frameworks (MOFs) are highly promising for carbon capture, yet navigating their vast design space remains challenging. Recent deep generative models enable de novo MOF design but primarily act as feed-forward structure generators. By heavily relying on predefined building block libraries and non-differentiable post-optimization, they fundamentally sever the information flow required for continuous structural editing. Here, we propose a target-driven generative framework focused on continuous structural manipulation. At its core is LinkerVAE, which maps discrete 3D chemical graphs into a continuous, SE(3)-equivariant latent space. This smooth manifold unlocks geometry-aware manipulations, including implicit chemical style transfer and zero-shot isoreticular expansion. Building upon this, we introduce a test-time optimization (TTO) strategy, utilizing an accurate surrogate model to continuously optimize the latent graphs of existing MOFs toward desired properties. This approach systematically enhances carbon capture performance, achieving a striking average relative boost of 147.5% in pure CO2 uptake while strictly preserving structural validity. Integrated with a latent diffusion model and rigid-body assembly for full MOF construction, our framework establishes a scalable, fully differentiable pathway for both the automated discovery, targeted optimization and editing of functional materials.

cs.LG↗

arXiv2026

AV-Unified: A Unified Framework for Audio-visual Scene Understanding

When humans perceive the world, they naturally integrate multiple audio-visual tasks within dynamic, real-world scenes. However, current works such as event localization, parsing, segmentation and question answering are mostly explored individually, making it challenging to comprehensively understand complex audio-visual scenes and explore inter-task relationships. Hence, we propose \textbf{AV-Unified}, a unified framework that enables joint learning across a wide range of audio-visual scene understanding tasks. AV-Unified standardizes the diverse input-output formats of each task and incorporates a multi-scale spatiotemporal perception network to effectively capture audio-visual associations. Specifically, we unify the inputs and outputs of all supported tasks by converting them into sequences of discrete tokens, establishing a shared representation that allows a single architecture to be trained jointly across heterogeneous varied datasets. Considering the varying temporal granularity of audio-visual events, a multi-scale temporal perception module is designed to capture key cues. Meanwhile, to overcome the lack of auditory supervision in the visual domain, we design a cross-modal guidance-based spatial perception module that models spatial audio-visual associations. Furthermore, task-specific text prompts are employed to enhance the model's adaptability and task-awareness. Extensive experiments on benchmark datasets (e.g., AVE, LLP, MUSIC-AVQA, VGG-SS and AVS) demonstrate the effectiveness of AV-Unified across temporal, spatial, and spatiotemporal tasks.

cs.CV↗

arXiv2026

Geodesic Gradient Descent: A Generic and Learning-rate-free Optimizer on Objective Function-induced Manifolds

Euclidean gradient descent algorithms barely capture the geometry of objective function-induced hypersurfaces and risk driving update trajectories off the hypersurfaces. Riemannian gradient descent algorithms address these issues but fail to represent complex hypersurfaces via a single classic manifold. We propose geodesic gradient descent (GGD), a generic and learning-rate-free Riemannian gradient descent algorithm. At each iteration, GGD uses an n-dimensional sphere to approximate a local neighborhood on the objective function-induced hypersurface, adapting to arbitrarily complex geometries. A tangent vector derived from the Euclidean gradient is projected onto the sphere to form a geodesic, ensuring the update trajectory stays on the hypersurface. Parameter updates are performed using the endpoint of the geodesic. The maximum step size of the gradient in GGD is equal to a quarter of the arc length on the n-dimensional sphere, thus eliminating the need for a learning rate. Experimental results show that compared with the classic Adam algorithm, GGD achieves test MSE reductions ranging from 35.79% to 48.76% for fully connected networks on the Burgers' dataset, and cross-entropy loss reductions ranging from 3.14% to 11.59% for convolutional neural networks on the MNIST dataset.

cs.LG↗

arXiv2026

Think with Grounding: Curriculum Reinforced Reasoning with Video Grounding for Long Video Understanding

Long video understanding is challenging due to rich and complicated multimodal clues in long temporal range.Current methods adopt reasoning to improve the model's ability to analyze complex video clues in long videos via text-form reasoning.However,the existing literature suffers from the fact that the text-only reasoning under fixed video context may exacerbate hallucinations since detailed crucial clues are often ignored under limited video context length due to the temporal redundancy of long videos.To address this gap,we propose Video-TwG,a curriculum reinforced framework that employs a novel Think-with-Grounding paradigm,enabling video LLMs to actively decide when to perform on-demand grounding during interleaved text-video reasoning, selectively zooming into question-relevant clips only when necessary.Video-TwG can be trained end-to-end in a straightforward manner, without relying on complex auxiliary modules or heavily annotated reasoning tracesIn detail,we design a Two-stage Reinforced Curriculum Strategy, where the model first learns think-with-grounding behavior on a small short-video GQA dataset with grounding labels,and then scales to diverse general QA data with videos of diverse domains to encourage generalization. Further, to handle complex think-with-grounding reasoning for various kinds of data,we propose TwG-GRPO algorithm which features the fine-grained grounding reward, self-confirmed pseudo reward and accuracy-gated mechanism.Finally,we propose to construct a new TwG-51K dataset that facilitates training. Experiments on Video-MME, LongVideoBench, and MLVU show that Video-TwG consistently outperforms strong LVU baselines.Further ablation validates the necessity of our Two-stage Reinforced Curriculum Strategy and shows our TwG-GRPO better leverages diverse unlabeled data to improve grounding quality and reduce redundant groundings without sacrificing QA performance.

cs.CV↗

arXiv2025

PhyVLLM: Physics-Guided Video Language Model with Motion-Appearance Disentanglement

Video Large Language Models (Video LLMs) have shown impressive performance across a wide range of video-language tasks. However, they often fail in scenarios requiring a deeper understanding of physical dynamics. This limitation primarily arises from their reliance on appearance-based matching. Incorporating physical motion modeling is crucial for deeper video understanding, but presents three key challenges: (1) motion signals are often entangled with appearance variations, making it difficult to extract clean physical cues; (2) effective motion modeling requires not only continuous-time motion representations but also capturing physical dynamics; and (3) collecting accurate annotations for physical attributes is costly and often impractical. To address these issues, we propose PhyVLLM, a physical-guided video-language framework that explicitly incorporates physical motion into Video LLMs. Specifically, PhyVLLM disentangles visual appearance and object motion through a dual-branch encoder. To model physical dynamics over time, we incorporate a Neural Ordinary Differential Equation (Neural ODE) module, which generates differentiable physical dynamic representations. The resulting motion-aware representations are projected into the token space of a pretrained LLM, enabling physics reasoning without compromising the model's original multimodal capabilities. To circumvent the need for explicit physical labels, PhyVLLM employs a self-supervised manner to model the continuous evolution of object motion. Experimental results demonstrate that PhyVLLM significantly outperforms state-of-the-art Video LLMs on both physical reasoning and general video understanding tasks, highlighting the advantages of incorporating explicit physical modeling.

cs.CV↗

arXiv2025

Solving Distance-Based Optimization Problems Using Optical Hardware

We present a practical approach to solving distance-based optimization problems using optical computing hardware. The objective is to minimize an energy function defined as the weighted sum of squared differences between measured distances and the squared Euclidean distances of point coordinates. By representing coordinates as complex numbers, we map the optimization problem onto optical fields, enabling its solution through either the canonical transformation (CT) method or the gain-based bifurcation (GBB) method. To further enhance the performance of the CT method, we introduce two techniques: asynchronous update and steepened gradient. Both the CT and GBB methods can effectively solve the distance-based problem and are adaptable to various optical hardware platforms. Our optical implementation is inspired by recent progress in analog optical computing for combinatorial optimization, highlighting its promise for efficient and scalable problem-solving in high-dimensional settings.

physics.optics↗

arXiv2025

Mettle: Meta-Token Learning for Memory-Efficient Audio-Visual Adaptation

We present \textbf{Met}a-\textbf{T}oken \textbf{Le}arning (Mettle), a simple and memory-efficient method for adapting large-scale pretrained transformer models to downstream audio-visual tasks. Instead of sequentially modifying the output feature distribution of the transformer backbone, Mettle utilizes a lightweight \textit{Layer-Centric Distillation (LCD)} module to distill in parallel the intact audio or visual features embedded by each transformer layer into compact meta-tokens. This distillation process considers both pretrained knowledge preservation and task-specific adaptation. The obtained meta-tokens can be directly applied to classification tasks, such as audio-visual event localization and audio-visual video parsing. To further support fine-grained segmentation tasks, such as audio-visual segmentation, we introduce a \textit{Meta-Token Injection (MTI)} module, which utilizes the audio and visual meta-tokens distilled from the top transformer layer to guide feature adaptation in earlier layers. Extensive experiments on multiple audiovisual benchmarks demonstrate that our method significantly reduces memory usage and training time while maintaining parameter efficiency and competitive accuracy.

cs.CV↗

arXiv2025

Phase Retrieval via Gain-Based Photonic XY-Hamiltonian Optimization

Phase-retrieval from coded diffraction patterns (CDP) is important to X-ray crystallography, diffraction tomography and astronomical imaging, yet remains a hard, non-convex inverse problem. We show that CDP recovery can be reformulated exactly as the minimisation of a continuous-variable XY Hamiltonian and solved by gain-based photonic networks. The coupled-mode equations we exploit are the natural mean-field dynamics of exciton-polariton condensate lattices, coupled-laser arrays and driven photon Bose-Einstein condensates, while other hardware such as the spatial photonic Ising machine can implement the same update rule through high-speed digital feedback, preserving full optical parallelism. Numerical experiments on images, two- and three-dimensional vortices and unstructured complex data demonstrate that the gain-based solver consistently outperforms the state-of-the-art Relaxed-Reflect-Reflect (RRR) algorithm in the medium-noise regime (signal-to-noise ratios 10--40 dB) and retains this advantage as problem size scales. Because the physical platform performs the continuous optimisation, our approach promises fast, energy-efficient phase retrieval on readily available photonic hardware. uch as two- and three-dimensional vortices, and unstructured random data. Moreover, the solver's accuracy remains high as problem sizes increase, underscoring its scalability.

physics.optics↗

arXiv2025

Bitcoin, a DAO?

This paper investigates whether Bitcoin can be regarded as a decentralized autonomous organization (DAO), what insights it may offer for the broader DAO ecosystem, and how Bitcoin governance can be improved. First, a quantitative literature analysis reveals that Bitcoin is increasingly overlooked in DAO research, even though early works often classified it as a DAO. Next, the paper applies a DAO viability framework - centering on collective intelligence, digital democracy, and adaptation - to examine Bitcoin's organizational and governance mechanisms. Findings suggest that Bitcoin instantitates key DAO principles by enabling open participation, and employing decentralized decision-making through Bitcoin Improvement Proposals (BIPs), miner signaling, and user-activated soft forks. However, this governance carries potential risks, including reduced clarity on who truly 'votes' due to the concentration of economic power among large stakeholders. The paper concludes by highlighting opportunities to refine Bitcoin's deliberation process and reflecting on broader implications for DAO design, such as the absence of a legal entity. In doing so, it underscores Bitcoin's continued relevance as an archetype for decentralized governance, offering important findings for future DAO implementations.

cs.CY↗

arXiv2025

Crab: A Unified Audio-Visual Scene Understanding Model with Explicit Cooperation

In recent years, numerous tasks have been proposed to encourage model to develop specified capability in understanding audio-visual scene, primarily categorized into temporal localization, spatial localization, spatio-temporal reasoning, and pixel-level understanding. Instead, human possesses a unified understanding ability for diversified tasks. Therefore, designing an audio-visual model with general capability to unify these tasks is of great value. However, simply joint training for all tasks can lead to interference due to the heterogeneity of audiovisual data and complex relationship among tasks. We argue that this problem can be solved through explicit cooperation among tasks. To achieve this goal, we propose a unified learning method which achieves explicit inter-task cooperation from both the perspectives of data and model thoroughly. Specifically, considering the labels of existing datasets are simple words, we carefully refine these datasets and construct an Audio-Visual Unified Instruction-tuning dataset with Explicit reasoning process (AV-UIE), which clarifies the cooperative relationship among tasks. Subsequently, to facilitate concrete cooperation in learning stage, an interaction-aware LoRA structure with multiple LoRA heads is designed to learn different aspects of audiovisual data interaction. By unifying the explicit cooperation across the data and model aspect, our method not only surpasses existing unified audio-visual model on multiple tasks, but also outperforms most specialized models for certain tasks. Furthermore, we also visualize the process of explicit cooperation and surprisingly find that each LoRA head has certain audio-visual understanding ability. Code and dataset: https://github.com/GeWu-Lab/Crab

cs.CV↗

arXiv2025

EvolvingAgent: Curriculum Self-evolving Agent with Continual World Model for Long-Horizon Tasks

Completing Long-Horizon (LH) tasks in open-ended worlds is an important yet difficult problem for embodied agents. Existing approaches suffer from two key challenges: (1) they heavily rely on experiences obtained from human-created data or curricula, failing to autonomously update and select multimodal experiences, and (2) they may encounter catastrophic forgetting issues when faced with new tasks, failing to autonomously update world knowledge. To solve these challenges, this paper presents {\bf EvolvingAgent}, a curriculum self-evolving agent with a continual World Model (WM), which can autonomously complete various LH tasks across environments through self-planning, self-control, and self-reflection, without human intervention. Specifically, EvolvingAgent contains three modules, i.e., i) the experience-driven task planner, which uses an LLM along with multimodal experiences to convert LH tasks into executable sub-tasks; ii) the WM-guided action controller, which leverages WM to generate low-level actions and incorporates a self-verification mechanism to update multimodal experiences; iii) the Curriculum Learning (CL) -based reflector, which implements a two-stage CL algorithm to select multimodal experiences for task-adaptive WM updates. By building a planner-controller-reflector closed-loop dynamic, the continual WM for EvolvingAgent can autonomously update multimodal experiences and world knowledge. We conducted extensive experiments on Minecraft, compared with existing methods, EvolvingAgent can improve 111.74{\%} average success rate, reduce more than 6x ineffective actions, and generalize to the Atari environment with human-level performance.

cs.RO↗

arXiv2024

Adaptive Prompt Learning with SAM for Few-shot Scanning Probe Microscope Image Segmentation

The Segment Anything Model (SAM) has demonstrated strong performance in image segmentation of natural scene images. However, its effectiveness diminishes markedly when applied to specific scientific domains, such as Scanning Probe Microscope (SPM) images. This decline in accuracy can be attributed to the distinct data distribution and limited availability of the data inherent in the scientific images. On the other hand, the acquisition of adequate SPM datasets is both time-intensive and laborious as well as skill-dependent. To address these challenges, we propose an Adaptive Prompt Learning with SAM (APL-SAM) framework tailored for few-shot SPM image segmentation. Our approach incorporates two key innovations to enhance SAM: 1) An Adaptive Prompt Learning module leverages few-shot embeddings derived from limited support set to learn adaptively central representatives, serving as visual prompts. This innovation eliminates the need for time-consuming online user interactions for providing prompts, such as exhaustively marking points and bounding boxes slice by slice; 2) A multi-source, multi-level mask decoder specifically designed for few-shot SPM image segmentation is introduced, which can effectively capture the correspondence between the support and query images. To facilitate comprehensive training and evaluation, we introduce a new dataset, SPM-Seg, curated for SPM image segmentation. Extensive experiments on this dataset reveal that the proposed APL-SAM framework significantly outperforms the original SAM, achieving over a 30% improvement in terms of Dice Similarity Coefficient with only one-shot guidance. Moreover, APL-SAM surpasses state-of-the-art few-shot segmentation methods and even fully supervised approaches in performance. Code and dataset used in this study will be made available upon acceptance.

cs.CV↗

arXiv2024

Multi-weather Cross-view Geo-localization Using Denoising Diffusion Models

Cross-view geo-localization in GNSS-denied environments aims to determine an unknown location by matching drone-view images with the correct geo-tagged satellite-view images from a large gallery. Recent research shows that learning discriminative image representations under specific weather conditions can significantly enhance performance. However, the frequent occurrence of unseen extreme weather conditions hinders progress. This paper introduces MCGF, a Multi-weather Cross-view Geo-localization Framework designed to dynamically adapt to unseen weather conditions. MCGF establishes a joint optimization between image restoration and geo-localization using denoising diffusion models. For image restoration, MCGF incorporates a shared encoder and a lightweight restoration module to help the backbone eliminate weather-specific information. For geo-localization, MCGF uses EVA-02 as a backbone for feature extraction, with cross-entropy loss for training and cosine distance for testing. Extensive experiments on University160k-WX demonstrate that MCGF achieves competitive results for geo-localization in varying weather conditions.

cs.CV↗

arXiv2024

Boosting Audio Visual Question Answering via Key Semantic-Aware Cues

The Audio Visual Question Answering (AVQA) task aims to answer questions related to various visual objects, sounds, and their interactions in videos. Such naturally multimodal videos contain rich and complex dynamic audio-visual components, with only a portion of them closely related to the given questions. Hence, effectively perceiving audio-visual cues relevant to the given questions is crucial for correctly answering them. In this paper, we propose a Temporal-Spatial Perception Model (TSPM), which aims to empower the model to perceive key visual and auditory cues related to the questions. Specifically, considering the challenge of aligning non-declarative questions and visual representations into the same semantic space using visual-language pretrained models, we construct declarative sentence prompts derived from the question template, to assist the temporal perception module in better identifying critical segments relevant to the questions. Subsequently, a spatial perception module is designed to merge visual tokens from selected segments to highlight key latent targets, followed by cross-modal interaction with audio to perceive potential sound-aware areas. Finally, the significant temporal-spatial cues from these modules are integrated to answer the question. Extensive experiments on multiple AVQA benchmarks demonstrate that our framework excels not only in understanding audio-visual scenes but also in answering complex questions effectively. Code is available at https://github.com/GeWu-Lab/TSPM.

cs.CV↗

arXiv2024

Ref-AVS: Refer and Segment Objects in Audio-Visual Scenes

Traditional reference segmentation tasks have predominantly focused on silent visual scenes, neglecting the integral role of multimodal perception and interaction in human experiences. In this work, we introduce a novel task called Reference Audio-Visual Segmentation (Ref-AVS), which seeks to segment objects within the visual domain based on expressions containing multimodal cues. Such expressions are articulated in natural language forms but are enriched with multimodal cues, including audio and visual descriptions. To facilitate this research, we construct the first Ref-AVS benchmark, which provides pixel-level annotations for objects described in corresponding multimodal-cue expressions. To tackle the Ref-AVS task, we propose a new method that adequately utilizes multimodal cues to offer precise segmentation guidance. Finally, we conduct quantitative and qualitative experiments on three test subsets to compare our approach with existing methods from related tasks. The results demonstrate the effectiveness of our method, highlighting its capability to precisely segment objects using multimodal-cue expressions. Dataset is available at \href{https://gewu-lab.github.io/Ref-AVS}{https://gewu-lab.github.io/Ref-AVS}.

cs.CV↗

arXiv2024

Collective modes in an unconventional superconductor with $j=3/2$ fermions

The $j = 3/2$ fermions in cubic crystals or cold atomic gases can form Cooper pairs in both singlet ($J = 0$) and unconventional quintet ($J = 2$) $s$-wave states. Our study utilizes analytical field theory to examine fluctuations in these states within the framework of the Luttinger-Kohn model. We investigate how collective modes evolve with varying spin-orbit coupling (SOC) strength. In the singlet state, quintet Bardasis-Schrieffer modes soften at a finite wavevector, hinting at Fulde-Ferrell-Larkin-Ovchinnikov physics. In the quintet state, we identify additional gapless and gapped modes originating from the partially broken symmetry due to SOC. Our results can be readily detected using current experimental techniques.

cond-mat.supr-con↗

arXiv2023

Audio-Visual Instance Segmentation

In this paper, we propose a new multi-modal task, termed audio-visual instance segmentation (AVIS), which aims to simultaneously identify, segment and track individual sounding object instances in audible videos. To facilitate this research, we introduce a high-quality benchmark named AVISeg, containing over 90K instance masks from 26 semantic categories in 926 long videos. Additionally, we propose a strong baseline model for this task. Our model first localizes sound source within each frame, and condenses object-specific contexts into concise tokens. Then it builds long-range audio-visual dependencies between these tokens using window-based attention, and tracks sounding objects among the entire video sequences. Extensive experiments reveal that our method performs best on AVISeg, surpassing the existing methods from related tasks. We further conduct the evaluation on several multi-modal large models. Unfortunately, they exhibits subpar performance on instance-level sound source localization and temporal perception. We expect that AVIS will inspire the community towards a more comprehensive multi-modal understanding. Dataset and code is available at https://github.com/ruohaoguo/avis.

cs.CV↗