Weifeng Su

Su Wang, Seyyedali Hosseinalipour, Vaneet Aggarwal et al.

Federated learning (FL) has been promoted as a popular technique for training machine learning (ML) models over edge/fog networks. Traditional implementations of FL have largely neglected the potential for inter-network cooperation, treating edge/fog devices and other infrastructure participating in ML as separate processing elements. Consequently, FL has been vulnerable to several dimensions of network heterogeneity, such as varying computation capabilities, communication resources, data qualities, and privacy demands. We advocate for cooperative federated learning (CFL), a cooperative edge/fog ML paradigm built on device-to-device (D2D) and device-to-server (D2S) interactions. Through D2D and D2S cooperation, CFL counteracts network heterogeneity in edge/fog networks through enabling a model/data/resource pooling mechanism, which will yield substantial improvements in ML model training quality and network resource consumption. We propose a set of core methodologies that form the foundation of D2D and D2S cooperation and present preliminary experiments that demonstrate their benefits. We also discuss new FL functionalities enabled by this cooperative framework such as the integration of unlabeled data and heterogeneous device privacy into ML model training. Finally, we describe some open research directions at the intersection of cooperative edge/fog and FL.

Yuchen Guo, Junli Gong, Hongmin Cai et al.

Recent segmentation models couple large language models (LLMs) with mask decoders to ground complex language expressions into masks, yet their instructions remain target-referential: they describe, constrain, or imply the region to be segmented. However, in real-world embodied interaction, human instructions are often at the intent-level, which includes the desired outcome without naming the region that enables it. To bridge this gap, we introduce SegWorld, where the model reasons about the scene through a multi-level visual chain-of-thought (CoT) before committing to a mask. Before receiving any instructions, it proactively observes the scene, describing visible objects and inferring plausible events they may support. Given an instruction, it continues the chain: from the object relevant to the intent, through the action that satisfies it, to the physical interaction site, the object part that affords the action. We formalize SegWorld as probabilistic inference, in which proactive observation supplies a linguistic scene context that improves mask prediction when instructions are given at the level of intent. We construct an intent-to-part benchmark for evaluating affordance-bearing part segmentation from high-level goals. Experiments show SegWorld matches instruction-driven baselines on target-referential instructions and improves substantially on intent-level ones.

Yuchen Guo, Junli Gong, Hongmin Cai et al.

We present PEAM, a Parametric Embodied Agent Memory framework in Minecraft that transforms agent memory from inference-time retrieval into parameter-resident skills internalized through experience. PEAM pairs a slow deliberative LLM for open-ended reasoning with a fast parametric module for reflexive execution of consolidated skills. The fast module is a multimodal Mixture-of-Experts LoRA architecture with per-category physically isolated adapters, enabling parameter-level continual learning without catastrophic forgetting. We treat failure as a first-class training signal: failure--correction trajectory pairs are internalized through a joint behavioral-cloning and contrastive objective, so the agent learns not only what succeeds but also how corrected actions differ from failed ones. To govern consolidation, PEAM introduces a parameterization-worthiness score for deciding which experience should be internalized, and a scale-free self-triggered consolidation mechanism for deciding when to internalize without task-specific hand-tuned thresholds, making the agent self-evolving as the trigger transfers across task distributions without re-tuning. Experiments in Minecraft show that PEAM improves long-horizon task performance, mitigates forgetting on previously consolidated skills, and improves parametric-versus-retrieval efficiency over retrieval-based embodied agents and parametric memory variants.

Payam Abdisarabshali, Nicholas Accurso, Filippo Malandra et al.

Federated learning (FL) is a popular distributed machine learning (ML) technique in Internet of Things (IoT) networks, where resource-constrained devices collaboratively train ML models while preserving data privacy. However, implementation of FL over 5G-and-beyond wireless networks faces key challenges caused by (i) dynamics of the wireless network conditions and (ii) the coexistence of multiple FL-services in the system. In this paper, we unveil two key phenomena that arise from these challenges: over/under-provisioning of resources and perspective-driven load balancing, both of which significantly impact FL performance in IoT environments. We take the first steps towards addressing these phenomena by proposing a novel distributed ML architecture called elastic FL (EFL). EFL unleashes the full potential of Open RAN (O-RAN) systems and introduces an elastic resource provisioning methodology to execute FL-services. It further constitutes a multi-time-scale FL management system that introduces three dedicated network control functionalities tailored for FL-services, including (i) non-real-time (non-RT) system descriptor, which trains ML-based applications to predict both system and FL-related dynamics and parameters; (ii) near-RT FL controller, which handles O-RAN slicing and mobility management for the seamless execution of FL-services; (iii) FL MAC scheduler, which conducts real-time resource allocation to the end clients of various FL-services. We finally prototype EFL to demonstrate its potential in improving the performance of FL-services.

Xinze Li, Bohan Yang, Pengxu Chen et al.

3D Gaussian Splatting (3DGS) has emerged as an advanced technique for real-time novel view synthesis by representing scene geometry and appearance using differentiable Gaussian primitives. However, efficiently computing precise Gaussian-tile intersections remains a critical task in the rasterization pipeline. To this end, we propose QuadBox, a method that leverages four axis-aligned bounding boxes to tightly encapsulate projected Gaussians in a discrete manner. First, we derive a geometry-aware stretching factor that enables the construction of a tile-aligned QuadBox, which covers the elliptical projection and largely excludes irrelevant tiles. Second, we introduce QPass, a single-pass tile traversal algorithm that exhaustively exploits the discrete nature of QuadBox, ensuring that the tile intersection check is performed with simple interval tests. Experiments on public datasets show that our method accelerates the rendering speed of 3DGS by 1.85$\times$. Code is available at \href{https://github.com/Powertony102/QuadBox}{https://github.com/Powertony102/QuadBox}.

Xinze Li, Pengxu Chen, Yiyuan Wang et al.

Feed-forward 3D foundation models face a key challenge: the quadratic computational cost introduced by global attention, which severely limits scalability as input length increases. Concurrent acceleration methods, such as token merging, operate at the token level. While they offer local savings, the required nearest-neighbor searches introduce undesirable overhead. Consequently, these techniques fail to tackle the fundamental issue of structural redundancy dominant in dense capture data. In this work, we introduce \textbf{S-VGGT}, a novel approach that addresses redundancy at the structural frame level, drastically shifting the optimization focus. We first leverage the initial features to build a dense scene graph, which characterizes structural scene redundancy and guides the subsequent scene partitioning. Using this graph, we softly assign frames to a small number of subscenes, guaranteeing balanced groups and smooth geometric transitions. The core innovation lies in designing the subscenes to share a common reference frame, establishing a parallel geometric bridge that enables independent and highly efficient processing without explicit geometric alignment. This structural reorganization provides strong intrinsic acceleration by cutting the global attention cost at its source. Crucially, S-VGGT is entirely orthogonal to token-level acceleration methods, allowing the two to be seamlessly combined for compounded speedups without compromising reconstruction fidelity. Code is available at https://github.com/Powertony102/S-VGGT.

Mingjie Zhao, Zhanpei Huang, Yang Lu et al.

Categorical attributes with qualitative values are ubiquitous in cluster analysis of real datasets. Unlike the Euclidean distance of numerical attributes, the categorical attributes lack well-defined relationships of their possible values (also called categories interchangeably), which hampers the exploration of compact categorical data clusters. Although most attempts are made for developing appropriate distance metrics, they typically assume a fixed topological relationship between categories when learning distance metrics, which limits their adaptability to varying cluster structures and often leads to suboptimal clustering performance. This paper, therefore, breaks the intrinsic relationship tie of attribute categories and learns customized distance metrics suitable for flexibly and accurately revealing various cluster distributions. As a result, the fitting ability of the clustering algorithm is significantly enhanced, benefiting from the learnable category relationships. Moreover, the learned category relationships are proved to be Euclidean distance metric-compatible, enabling a seamless extension to mixed datasets that include both numerical and categorical attributes. Comparative experiments on 12 real benchmark datasets with significance tests show the superior clustering accuracy of the proposed method with an average ranking of 1.25, which is significantly higher than the 5.21 ranking of the current best-performing method.

Zhongyi Yu, Jianqiu Wu, Zhenghao Wu et al.

Temporal graph link prediction aims to predict future interactions between nodes in a graph based on their historical interactions, which are encoded in node embeddings. We observe that heterogeneity naturally appears in temporal interactions, e.g., a few node pairs can make most interaction events, and interaction events happen at varying intervals. This leads to the problems of ineffective temporal information encoding and forgetting of past interactions for a pair of nodes that interact intermittently for their link prediction. Existing methods, however, do not consider such heterogeneity in their learning process, and thus their learned temporal node embeddings are less effective, especially when predicting the links for infrequently interacting node pairs. To cope with the heterogeneity, we propose a novel framework called TAMI, which contains two effective components, namely log time encoding function (LTE) and link history aggregation (LHA). LTE better encodes the temporal information through transforming interaction intervals into more balanced ones, and LHA prevents the historical interactions for each target node pair from being forgotten. State-of-the-art temporal graph neural networks can be seamlessly and readily integrated into TAMI to improve their effectiveness. Experiment results on 13 classic datasets and three newest temporal graph benchmark (TGB) datasets show that TAMI consistently improves the link prediction performance of the underlying models in both transductive and inductive settings. Our code is available at https://github.com/Alleinx/TAMI_temporal_graph.

Yuqian Wu, Yuhong Peng, Jiapeng Yu et al.

Next location prediction is a key task in human mobility analysis, crucial for applications like smart city resource allocation and personalized navigation services. However, existing methods face two significant challenges: first, they fail to address the dynamic imbalance between periodic and chaotic mobile patterns, leading to inadequate adaptation over sparse trajectories; second, they underutilize contextual cues, such as temporal regularities in arrival times, which persist even in chaotic patterns and offer stronger predictability than spatial forecasts due to reduced search spaces. To tackle these challenges, we propose \textbf{\method}, a \underline{\textbf{C}}h\underline{\textbf{A}}otic \underline{\textbf{N}}eural \underline{\textbf{O}}scillator n\underline{\textbf{E}}twork for next location prediction, which introduces a biologically inspired Chaotic Neural Oscillatory Attention mechanism to inject adaptive variability into traditional attention, enabling balanced representation of evolving mobility behaviors, and employs a Tri-Pair Interaction Encoder along with a Cross Context Attentive Decoder to fuse multimodal ``who-when-where'' contexts in a joint framework for enhanced prediction performance. Extensive experiments on two real-world datasets demonstrate that CANOE consistently and significantly outperforms a sizeable collection of state-of-the-art baselines, yielding 3.17\%-13.11\% improvement over the best-performing baselines across different cases. In particular, CANOE can make robust predictions over mobility trajectories of different mobility chaotic levels. A series of ablation studies also supports our key design choices. Our code is available at: https://github.com/yuqian2003/CANOE.

Zhongyu Mai, Zewei Zhan, Hanyu Guo et al.

Magnetic resonance imaging (MRI) motion artifacts can seriously affect clinical diagnostics, making it challenging to interpret images accurately. Existing methods for eliminating motion artifacts struggle to retain fine structural details and simultaneously lack the necessary vividness and sharpness. In this study, we present a novel dual-domain optimization (DDO) approach that integrates information from the pixel and frequency domains guiding the recovery of clean magnetic resonance images through implicit neural representations(INRs). Specifically, our approach leverages the low-frequency components in the k-space as a reference to capture accurate tissue textures, while high-frequency and pixel information contribute to recover details. Furthermore, we design complementary masks and dynamic loss weighting transitioning from global to local attention that effectively suppress artifacts while retaining useful details for reconstruction. Experimental results on the NYU fastMRI dataset demonstrate that our method outperforms existing approaches in multiple evaluation metrics. Our code is available at https://anonymous.4open.science/r/DDO-IN-A73B.

Yuchen Guo, Junli Gong, Wenjun Dong et al.

Purely RGB-based vision models often fail to provide reliable cues in challenging scenarios such as nighttime and fog, leading to degraded performance and safety risks. Infrared imaging captures heat-emitting sources and provides critical complementary information, but existing high-fidelity fusion methods suffer from prohibitive latency, rendering them impractical for real-time edge deployment. To address this, we propose FusionProxy, a real-time image fusion module designed as a fully independent, plug-and-play component with diffusion level quality. FusionProxy exploits two complementary statistics of a teacher sample ensemble: per-pixel variance in raw image space, used to weight pixel-level supervision, and per-pixel variance inside frozen foundation backbones, used to route feature-level alignment spatially. Once trained, FusionProxy can be directly integrated into any visual perception system without joint optimization. Extensive experiments demonstrate that our method achieves superior performance on static recognition tasks and significantly enhances robustness in dynamic tasks, including closed-loop autonomous driving. Crucially, FusionProxy achieves real-time inference speeds on diverse platforms, from high-end GPUs to commodity hardware, providing a flexible and generalizable solution for all-day perception.

Yuchen Guo, Junli Gong, Yao Lu et al.

Infrared-Visible image fusion (IVIF) aims to integrate thermal information and detailed spatial structures into a single fused image to enhance perception. However, existing evaluation approaches tend to over-optimize both hand-crafted no-reference statistics and full-reference metrics that treat the source images as pseudo ground truths. Recent IVIF reward-modelling efforts learn from human ratings but use scalar regression on aggregated scores, neither leveraging the reasoning of Multimodal Large Language Models (MLLMs) nor encoding per-image perceptual ambiguity in their supervision, but naively introducing MLLMs with discrete one-hot supervision likewise collapses fused images of similar quality into different rating levels. To address this, we introduce FuScore, which utilizes an MLLM to mimic human visual perception by producing continuous quality score, rather than discrete level predictions, enabling fine-grained discrimination among fused images of similar quality. We exploit the agreement among four IVIF-specific sub-dimensions to construct a per-image soft label whose sharpness reflects how consensual the overall judgment is. We further introduce a tripartite objective combining per-image distributional supervision, within-source-pair Thurstone fidelity for method-level ordering, and cross-source-pair Thurstone fidelity for scene-level ordering across scenes. Extensive experiments demonstrate that FuScore achieves state-of-the-art correlation with human visual preferences.

Yuchen Guo, Ruoxiang Xu, Rongcheng Li et al.

In extreme scenarios such as nighttime or low-visibility environments, achieving reliable perception is critical for applications like autonomous driving, robotics, and surveillance. Multi-modality image fusion, particularly integrating infrared imaging, offers a robust solution by combining complementary information from different modalities to enhance scene understanding and decision-making. However, current methods face significant limitations: GAN-based approaches often produce blurry images that lack fine-grained details, while AE-based methods may introduce bias toward specific modalities, leading to unnatural fusion results. To address these challenges, we propose DAE-Fuse, a novel two-phase discriminative autoencoder framework that generates sharp and natural fused images. Furthermore, We pioneer the extension of image fusion techniques from static images to the video domain while preserving temporal consistency across frames, thus advancing the perceptual capabilities required for autonomous navigation. Extensive experiments on public datasets demonstrate that DAE-Fuse achieves state-of-the-art performance on multiple benchmarks, with superior generalizability to tasks like medical image fusion.

Yuchen Guo, Weifeng Su

Although multi-modality medical image segmentation holds significant potential for enhancing the diagnosis and understanding of complex diseases by integrating diverse imaging modalities, existing methods predominantly rely on feature-level fusion strategies. We argue the current feature-level fusion strategy is prone to semantic inconsistencies and misalignments across various imaging modalities because it merges features at intermediate layers in a neural network without evaluative control. To mitigate this, we introduce a novel image-level fusion based multi-modality medical image segmentation method, Fuse4Seg, which is a bi-level learning framework designed to model the intertwined dependencies between medical image segmentation and medical image fusion. The image-level fusion process is seamlessly employed to guide and enhance the segmentation results through a layered optimization approach. Besides, the knowledge gained from the segmentation module can effectively enhance the fusion module. This ensures that the resultant fused image is a coherent representation that accurately amalgamates information from all modalities. Moreover, we construct a BraTS-Fuse benchmark based on BraTS dataset, which includes 2040 paired original images, multi-modal fusion images, and ground truth. This benchmark not only serves image-level medical segmentation but is also the largest dataset for medical image fusion to date. Extensive experiments on several public datasets and our benchmark demonstrate the superiority of our approach over prior state-of-the-art (SOTA) methodologies.

Payam Abdisarabshali, Kwang Taik Kim, Michael Langberg et al.

Existing studies on federated learning (FL) are mostly focused on system orchestration for static snapshots of the network and making static control decisions (e.g., spectrum allocation). However, real-world wireless networks are susceptible to temporal variations of wireless channel capacity and users' datasets. In this paper, we incorporate multi-granular system dynamics (MSDs) into FL, including (M1) dynamic wireless channel capacity, captured by a set of discrete-time events, called $\mathscr{D}$-Events, and (M2) dynamic datasets of users. The latter is characterized by (M2-a) modeling the dynamics of user's dataset size via an ordinary differential equation and (M2-b) introducing dynamic model drift}, formulated via a partial differential inequality} drawing concrete analytical connections between the dynamics of users' datasets and FL accuracy. We then conduct FL orchestration under MSDs by introducing dynamic cooperative FL with dedicated MAC schedulers (DCLM), exploiting the unique features of open radio access network (O-RAN). DCLM proposes (i) a hierarchical device-to-device (D2D)-assisted model training, (ii) dynamic control decisions through dedicated O-RAN MAC schedulers, and (iii) asymmetric user selection. We provide extensive theoretical analysis to study the convergence of DCLM. We then optimize the degrees of freedom (e.g., user selection and spectrum allocation) in DCLM through a highly non-convex optimization problem. We develop a systematic approach to obtain the solution for this problem, opening the door to solving a broad variety of network-aware FL optimization problems. We show the efficiency of DCLM via numerical simulations and provide a series of future directions.

Yuchen Guo, Junli Gong, Hongmin Cai et al.

Video color grading is a critical post-production process that transforms flat, log-encoded raw footage into emotionally resonant cinematic visuals. Existing automated methods act as static, black-box executors that directly output edited pixels, lacking both interpretability and the iterative control required by professionals. We introduce LumiVideo, an agentic system that mimics the cognitive workflow of professional colorists through four stages: Perception, Reasoning, Execution, and Reflection. Given only raw log video, LumiVideo autonomously produces a cinematic base grade by analyzing the scene's physical lighting and semantic content. Its Reasoning engine synergizes an LLM's internalized cinematic knowledge with a Retrieval-Augmented Generation (RAG) framework via a Tree of Thoughts (ToT) search to navigate the non-linear color parameter space. Rather than generating pixels, the system compiles the deduced parameters into industry-standard ASC-CDL configurations and a globally consistent 3D LUT, analytically guaranteeing temporal consistency. An optional Reflection loop then allows creators to refine the result via natural language feedback. We further introduce LumiGrade, the first log-encoded video benchmark for evaluating automated grading. Experiments show that LumiVideo approaches human expert quality in fully automatic mode while enabling precise iterative control when directed.

Xinze Li, Runlin Huang, Zhenghao Wu et al.

Semi-supervised learning (SSL) has been widely used to learn from both a few labeled images and many unlabeled images to overcome the scarcity of labeled samples in medical image segmentation. Most current SSL-based segmentation methods use pixel values directly to identify similar features in labeled and unlabeled data. They usually fail to accurately capture the intricate attachment structures in the left atrium, such as the areas of inconsistent density or exhibit outward curvatures, adding to the complexity of the task. In this paper, we delve into this issue and introduce an effective solution, CORAL(Correlation-Aligned)-Correlation Consistency Network (CORN), to capture the global structure shape and local details of Left Atrium. Diverging from previous methods focused on each local pixel value, the CORAL-Correlation Consistency Module (CCM) in the CORN leverages second-order statistical information to capture global structural features by minimizing the distribution discrepancy between labeled and unlabeled samples in feature space. Yet, direct construction of features from unlabeled data frequently results in ``Sample Selection Bias'', leading to flawed supervision. We thus further propose the Dynamic Feature Pool (DFP) for the CCM, which utilizes a confidence-based filtering strategy to remove incorrectly selected features and regularize both teacher and student models by constraining the similarity matrix to be consistent. Extensive experiments on the Left Atrium dataset have shown that the proposed CORN outperforms previous state-of-the-art semi-supervised learning methods.

Ruilang Wang, Shuotong Xu, Bowen Liu et al.

The scarcity of annotated data in specialized domains such as medical imaging presents significant challenges to training robust vision models. While self-supervised masked image modeling (MIM) offers a promising solution, existing approaches largely rely on random high-ratio masking, leading to inefficiency and poor semantic alignment. Moreover, region-aware variants typically depend on reconstruction heuristics or supervised signals, limiting their adaptability across tasks and modalities. We propose Mask What Matters, a controllable text-guided masking framework for self-supervised medical image analysis. By leveraging vision-language models for prompt-based region localization, our method flexibly applies differentiated masking to emphasize diagnostically relevant regions while reducing redundancy in background areas. This controllable design enables better semantic alignment, improved representation learning, and stronger cross-task generalizability. Comprehensive evaluation across multiple medical imaging modalities, including brain MRI, chest CT, and lung X-ray, shows that Mask What Matters consistently outperforms existing MIM methods (e.g., SparK), achieving gains of up to +3.1 percentage points in classification accuracy, +1.3 in box average precision (BoxAP), and +1.1 in mask average precision (MaskAP) for detection. Notably, it achieves these improvements with substantially lower overall masking ratios (e.g., 40\% vs. 70\%). This work demonstrates that controllable, text-driven masking can enable semantically aligned self-supervised learning, advancing the development of robust vision models for medical image analysis.

Zhongyi Yu, Zhenghao Wu, Shuhan Zhong et al.

Missing values are a common problem that poses significant challenges to data analysis and machine learning. This problem necessitates the development of an effective imputation method to fill in the missing values accurately, thereby enhancing the overall quality and utility of the datasets. Existing imputation methods, however, fall short of explicitly considering the `missingness' information in the data during the embedding initialization stage and modeling the entangled feature and sample correlations during the learning process, thus leading to inferior performance. We propose M$^3$-Impute, which aims to explicitly leverage the missingness information and such correlations with novel masking schemes. M$^3$-Impute first models the data as a bipartite graph and uses a graph neural network to learn node embeddings, where the refined embedding initialization process directly incorporates the missingness information. They are then optimized through M$^3$-Impute's novel feature correlation unit (FRU) and sample correlation unit (SRU) that effectively captures feature and sample correlations for imputation. Experiment results on 25 benchmark datasets under three different missingness settings show the effectiveness of M$^3$-Impute by achieving 20 best and 4 second-best MAE scores on average.

Zhongyi Yu, Zhenghao Wu, Hao Zheng et al.

Fixed length summarization aims at generating summaries with a preset number of words or characters. Most recent researches incorporate length information with word embeddings as the input to the recurrent decoding unit, causing a compromise between length controllability and summary quality. In this work, we present an effective length controlling unit Length Attention (LenAtten) to break this trade-off. Experimental results show that LenAtten not only brings improvements in length controllability and ROGUE scores but also has great generalization ability. In the task of generating a summary with the target length, our model is 732 times better than the best-performing length controllable summarizer in length controllability on the CNN/Daily Mail dataset.