Manifold-Based Verbalizer Space Re-embedding for Tuning-Free Prompt-Based Classification.
Haochun WangSendong ZhaoChi LiuNuwa XiMuzhen CaiBing QinTing LiuPublished in: AAAI (2024)
Keyphrases
- low dimensional
- feature space
- image classification
- pattern recognition
- classification accuracy
- vector space
- feature extraction
- geodesic distance
- embedding space
- feature vectors
- search space
- graph embedding
- text classification
- support vector
- classification algorithm
- parameter space
- laplacian eigenmaps
- decision trees
- latent space
- classification scheme
- manifold embedding
- manifold learning
- class labels
- training samples
- supervised learning
- support vector machine
- high dimensional
- support vector machine svm
- lower dimensional
- nonlinear dimensionality reduction
- training set