A5100583711- Text and Document Classification Technologies
- Image Retrieval and Classification Techniques
- Face and Expression Recognition
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Spam and Phishing Detection
- Machine Learning and ELM
- Machine Learning in Bioinformatics
Jiangxi Agricultural University
2023-2024
Partial multilabel learning (PML) addresses the issue of noisy supervision, which contains an overcomplete set candidate labels for each instance with only a valid subset training data. Using label enhancement techniques, researchers have computed probability being ground truth. However, enhancing in space makes it impossible existing partial methods to achieve satisfactory results. Besides, few simultaneously involve ambiguity problem, feature space's redundancy, and model's efficiency PML....
Partial Multi-Labeling Learning (PML) is a more practical learning paradigm, in which the labeling information ambiguated. Most existing PML algorithms rely on assumptions to resolve ambiguity. However, these do not account for origin of noise and therefore fail address impact learner's performance at root. In this paper, we will propose method jointly granular ball-based robust feature selection relevance fusion optimization (PML-GR). Specifically, first stage, construct ball compute...