entropy minimization
A training strategy that encourages the model to output high-confidence predictions by minimizing uncertainty in the predictions it makes, often used in semi-supervised and self-supervised learning.
- Active Test-time Vision-Language Navigation
- CLIPTTA: Robust Contrastive Vision-Language Test-Time Adaptation
- Decoupled Entropy Minimization
- Hierarchical Self-Attention: Generalizing Neural Attention Mechanics to Multi-Scale Problems
- Rethinking Entropy in Test-Time Adaptation: The Missing Piece from Energy Duality
- Test-Time Adaptation of Vision-Language Models for Open-Vocabulary Semantic Segmentation
- Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
- The Unreasonable Effectiveness of Entropy Minimization in LLM Reasoning