Bridging Generative and Discriminative Noisy-Label Learning via Direction-Agnostic EM Formulation
cs.CV, cs.LG
Submitted: 2023-08-02
Updated: 2026-09-26
Code: https://github.com/lfb-1/GNL
Terminology
Sources
- BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
- Causal Direction of Data Collection Matters: Implications of Causal and Anticausal Learning for NLP
- Blind Knowledge Distillation for Robust Image Classification
- Auto-Encoding Variational Bayes
- WebVision Database: Visual Learning and Understanding from Web Data
- Provably End-to-end Label-Noise Learning without Anchor Points
- Identifiability of Label Noise Transition Matrix
- Long-tail learning via logit adjustment
- Source-free Domain Adaptation via Avatar Prototype Generation and Adaptation
- Resolving label uncertainty with implicit posterior models
- To Smooth or Not? When Label Smoothing Meets Noisy Labels
- mixup: Beyond Empirical Risk Minimization
- Learning with Feature-Dependent Label Noise: A Progressive Approach
- Learning from Noisy Labels for Entity-Centric Information Extraction
- DyGen: Learning from Noisy Labels via Dynamics-Enhanced Generative Modeling
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models