Furao Shen

Learning More by Training Less: Uncertainty-Guided Progressive Contraction for Efficient Reasoning Post-Training

Suorong Yang, Jiaxin Wang, Jiaming Su +4

Large language models have achieved remarkable advances in complex reasoning. However, scaling reasoning post-training with ever-larger corpora can substantially increase training…

FedLAFP: Low-Rank Aggregation Meets Full-Rank Personalization in Federated Fine-Tuning

Mengjun Yi, Huaian Gu, Yinghao Ai +2

Federated parameter-efficient fine-tuning enables clients to adapt pre-trained models without sharing raw data or communicating the full model, but statistical heterogeneity makes…

CF-LoRA: Decoupled Factor Aggregation and Adaptation-Aware Client Clustering for Federated LoRA Fine-Tuning

Mengjun Yi, Langxing Yang, Suhan Guo +2

Federated LoRA fine-tuning enables parameter-efficient adaptation of pre-trained models without sharing private data, but suffers from two fundamental mismatches under heterogeneou…

Task-Anchored Representation Shaping for Pre-Trained Model-Based Continual Learning

Zhiming Xu, Huiyu Yi, Zhen-Hao Xie +4

Pre-trained models (PTMs) provide a strong foundation for continual learning by offering stable representations that facilitate lightweight adaptation to new tasks. However, adapti…

Beyond Routing Saturation: A Long-Horizon Class-Incremental Perspective on Expert Routing in Multimodal Continual Instruction Tuning

Huiyu Yi, Yongqi Xu, Bogang Zhang +5

Multimodal Continual Instruction Tuning (MCIT) enables multimodal large language models to acquire new tasks sequentially while retaining previously learned capabilities. Many rece…

cs.CV2026

Redirecting the Flow: Image Customization through Attention Distribution Shift

Jie Li, Suorong Yang, Jian Zhao +1

Subject-driven image customization aims to generate images that not only follow textual instructions but also preserve the identity of a given reference subject. Existing approache…

Directional Linear Separability of Neural Representations: Geometry and Transformations

Yi Wei, Xuan Qi, Furao Shen +1

Neural networks build representations through affine maps and nonlinear activations. Injective affine maps preserve linear separability, raising the problem of how they prepare dat…

Beyond What to Select: A Plug-and-play Oscillatory Data-Volume Scheduling for Efficient Model Training

Suorong Yang, Hanqi Zhu, Hai Gan +4

Data selection accelerates training by identifying representative training data while preserving model performance. However, existing methods mainly focus on designing sample-impor…

cs.CV2026

Beyond Point-wise Neural Collapse: A Topology-Aware Hierarchical Classifier for Class-Incremental Learning

Huiyu Yi, Zhiming Xu, Dunwei Tu +3

The Nearest Class Mean (NCM) classifier is widely favored in Class-Incremental Learning (CIL) for its superior resistance to catastrophic forgetting compared to Fully Connected lay…

AffineLens: Capturing the Continuous Piecewise Affine Functions of Neural Networks

Yi Wei, Xuan Qi, Furao Shen +3

Piecewise affine neural networks (PANNs) provide a principled geometric perspective on neural network expressivity by characterizing the input--output map as a continuous piecewise…

Training-Time Batch Normalization Reshapes Local Partition Geometry in Piecewise-Affine Networks

Xuan Qi, Yi Wei, Fanqi Yu +3

Batch normalization (BN) is central to modern deep networks, but its effect on the realized function during training remains less understood than its optimization benefits. We stud…

Region Seeding via Pre-Activation Regularization: A Geometric View of Piecewise Affine Neural Networks

Yi Wei, Xuan Qi, Furao Shen

Deep networks with continuous piecewise affine activations induce polyhedral partitions of the input space, making the number of realized affine regions a natural measure of expres…

Free-Flow Class-Incremental Learning: Towards Robust CIL under Variable Class Arrivals

Zhiming Xu, Baile Xu, Jian Zhao +2

Class-incremental learning (CIL) is commonly evaluated under predefined schedules with fixed or nearly equal class increments, leaving irregular class-arrival scenarios underexplor…

Data Agent: Learning to Select Data via End-to-End Dynamic Optimization

Suorong Yang, Fangjian Su, Hai Gan +5

Dynamic Data selection aims to accelerate training by prioritizing informative samples during online training. However, existing methods typically rely on task-specific handcrafted…

cs.CV2026

Disentangle to create: Architecture-level disentangled diffusion model for few-shot Chinese font generation

Jie Li, Suorong Yang, Jian Zhao +1

Few-shot Chinese font generation aims to synthesize new characters in a target style using only a handful of reference images. Achieving accurate content rendering and faithful sty…

T-LLM: Teaching Large Language Models to Forecast Time Series via Temporal Distillation

Suhan Guo, Bingxu Wang, Shaodan Zhang +1

Time series forecasting plays a critical role in decision-making across many real-world applications. Unlike data in vision and language domains, time series data is inherently tie…

MiCA: A Mobility-Informed Causal Adapter for Lightweight Epidemic Forecasting

Suhan Guo, Jiahong Deng, Furao Shen

Accurate forecasting of infectious disease dynamics is critical for public health planning and intervention. Human mobility plays a central role in shaping the spatial spread of ep…

Pushing the Limits of Distillation-Based Continual Learning via Classifier-Proximal Lightweight Plugins

Zhiming Xu, Baile Xu, Jian Zhao +2

Continual learning requires models to learn continuously while preserving prior knowledge under evolving data streams. Distillation-based methods are appealing for retaining past k…

On-the-Fly Data Augmentation via Gradient-Guided and Sample-Aware Influence Estimation

Suorong Yang, Jie Zong, Lihang Wang +6

Data augmentation has been widely employed to improve the generalization of deep neural networks. Most existing methods apply fixed or random transformations. However, we find that…

cs.CV2025

IPF-RDA: An Information-Preserving Framework for Robust Data Augmentation

Suorong Yang, Hongchao Yang, Suhan Guo +2

Data augmentation is widely utilized as an effective technique to enhance the generalization performance of deep models. However, data augmentation may inevitably introduce distrib…

Multimodal-Guided Dynamic Dataset Pruning for Robust and Efficient Data-Centric Learning

Suorong Yang, Peijia Li, Yujie Liu +5

Modern deep models are trained on large real-world datasets, where data quality varies and redundancy is common. Data-centric approaches such as dataset pruning have shown promise…

cs.DC2025

CacheFL: Privacy-Preserving and Efficient Federated Cache Model Fine-Tuning for Vision-Language Models

Mengjun Yi, Hanwen Zhang, Hui Dou +2

Large pre-trained Vision-Language Models (VLMs), such as Contrastive Language-Image Pre-training (CLIP), have exhibited remarkable zero-shot performance across various image classi…

Enhancing Epidemic Forecasting: Evaluating the Role of Mobility Data and Graph Convolutional Networks

Suhan Guo, Zhenghao Xu, Furao Shen +1

Accurate prediction of contagious disease outbreaks is vital for informed decision-making. Our study addresses the gap between machine learning algorithms and their epidemiological…