Tong Wei

Associate Professor@Southeast University·中国·OpenReview

研究方向

semi-supervised learning · learning with noisy labels · long-tailed learning · out-of-distribution detection · weakly-supervised learning · continual learning

5.5

KeepLoRA: Continual Learning with Residual Gradient Adaptation

ICLR 2026Poster

通讯

4.7

Unlabeled Data vs. Pre-trained Knowledge: Rethinking SSL in the Era of Large Models

ICLR 2026Rejected

三作

4.5

Semi-Supervised CLIP Adaptation by Enforcing Semantic and Trapezoidal Consistency

ICLR 2025Poster

通讯

5.5

LADA: Scalable Label-Specific CLIP Adapter for Continual Learning

ICML 2025Poster

三作

5.3

Harnessing Shallow Features in Pre-Trained Models for Out-of-Distribution Detection

合作者 (20)

Tong Wei

KeepLoRA: Continual Learning with Residual Gradient Adaptation

Unlabeled Data vs. Pre-trained Knowledge: Rethinking SSL in the Era of Large Models

How Does Fine-Tuned Foundation Models Help for Long-Tailed Data

Memory-Efficient Fine-Tuning via Low-Rank Activation Compression

Weakly-Supervised Contrastive Learning for Imprecise Class Labels

X-Mahalanobis: Transformer Feature Mixing for Reliable OOD Detection

Semi-Supervised CLIP Adaptation by Enforcing Semantic and Trapezoidal Consistency

LADA: Scalable Label-Specific CLIP Adapter for Continual Learning

Harnessing Shallow Features in Pre-Trained Models for Out-of-Distribution Detection