KI-Glossar
Das vollständige Wörterbuch der Künstlichen Intelligenz
Fine-Tuning
Technique for adjusting the weights of a pre-trained model on new data specific to the target task
Feature Extraction
Using the lower layers of a pre-trained model as a fixed feature extractor
Domain Adaptation
Adaptation of a model trained on a source domain to a target domain with a different distribution
Few-Shot Learning
Learning with very few training examples per class thanks to transferred knowledge
Zero-Shot Learning
Classification of never-seen objects during training by transferring semantic knowledge
Multi-Task Transfer
Sharing representations between multiple related tasks to improve overall performance
Cross-Lingual Transfer
Knowledge transfer between models trained on different languages
Progressive Networks
Architecture that preserves knowledge from previous tasks while learning new tasks
Knowledge Distillation
Compression of a large pre-trained model to a smaller model while preserving performance
Adversarial Transfer Learning
Using adversarial networks to align distributions between source and target domains
Transfer Learning for NLP
Specific application of transfer learning to natural language processing models like BERT and GPT
Transfer Learning for Computer Vision
Reuse of pre-trained vision models (ResNet, VGG) for specific image classification tasks
Meta-Learning
Learning to learn by optimizing the ability to quickly transfer to new tasks
Self-Supervised Transfer
Transfer from pre-trained models with self-supervision on massive unlabeled data