Výsledky vyhledávání - "Fan, Yunfeng"

Report

Detached and Interactive Multimodal Learning

Autor: Fan, Yunfeng, Xu, Wenchao, Wang, Haozhao, Liu, Junhong, Guo, Song

Recently, Multimodal Learning (MML) has gained significant interest as it compensates for single-modality limitations through comprehensive complementary information within multimodal data. However, traditional MML methods generally use the joint lea

Externí odkaz: http://arxiv.org/abs/2407.19514

Zobrazit plný text záznamu

Report

Overcome Modal Bias in Multi-modal Federated Learning via Balanced Modality Selection

Autor: Fan, Yunfeng, Xu, Wenchao, Wang, Haozhao, Huo, Fushuo, Chen, Jinyu, Guo, Song

Selecting proper clients to participate in each federated learning (FL) round is critical to effectively harness a broad range of distributed data. Existing client selection methods simply consider the mining of distributed uni-modal data, yet, their

Externí odkaz: http://arxiv.org/abs/2401.00403

Zobrazit plný text záznamu

Report

Balanced Multi-modal Federated Learning via Cross-Modal Infiltration

Autor: Fan, Yunfeng, Xu, Wenchao, Wang, Haozhao, Zhu, Jiaqi, Guo, Song

Federated learning (FL) underpins advancements in privacy-preserving distributed computing by collaboratively training neural networks without exposing clients' raw data. Current FL paradigms primarily focus on uni-modal data, while exploiting the kn

Externí odkaz: http://arxiv.org/abs/2401.00894

Zobrazit plný text záznamu

Report

Non-Exemplar Online Class-incremental Continual Learning via Dual-prototype Self-augment and Refinement

Autor: Huo, Fushuo, Xu, Wenchao, Guo, Jingcai, Wang, Haozhao, Fan, Yunfeng, Guo, Song

This paper investigates a new, practical, but challenging problem named Non-exemplar Online Class-incremental continual Learning (NO-CL), which aims to preserve the discernibility of base classes without buffering data examples and efficiently learn

Externí odkaz: http://arxiv.org/abs/2303.10891

Zobrazit plný text záznamu

Report

DualMix: Unleashing the Potential of Data Augmentation for Online Class-Incremental Learning

Autor: Fan, Yunfeng, Xu, Wenchao, Wang, Haozhao, Zhu, Jiaqi, Wang, Junxiao, Guo, Song

Online Class-Incremental (OCI) learning has sparked new approaches to expand the previously trained model knowledge from sequentially arriving data streams with new classes. Unfortunately, OCI learning can suffer from catastrophic forgetting (CF) as

Externí odkaz: http://arxiv.org/abs/2303.07864

Zobrazit plný text záznamu

Report

PMR: Prototypical Modal Rebalance for Multimodal Learning

Autor: Fan, Yunfeng, Xu, Wenchao, Wang, Haozhao, Wang, Junxiao, Guo, Song

Multimodal learning (MML) aims to jointly exploit the common priors of different modalities to compensate for their inherent limitations. However, existing MML methods often optimize a uniform objective for different modalities, leading to the notori

Externí odkaz: http://arxiv.org/abs/2211.07089

Zobrazit plný text záznamu