Порівняння методів
Переглядайте обрані методи поруч; рядки з відмінностями підсвічено.
| Ансамблеве трансферне навчання× | Навчання з малою кількістю прикладів× | |
|---|---|---|
| Галузь | Машинне навчання | Машинне навчання |
| Родина | Machine learning | Machine learning |
| Рік появи≠ | 2010s | 2011–2017 |
| Автор методу≠ | Various (consolidated in deep learning era, 2010s) | Lake, B. M.; Vinyals, O.; Finn, C. et al. |
| Тип≠ | Ensemble of pre-trained / fine-tuned models | Meta-learning / low-data learning paradigm |
| Основоположне джерело≠ | Ganaie, M. A., Hu, M., Malik, A. K., Tanveer, M., & Suganthan, P. N. (2022). Ensemble deep learning: A review. Engineering Applications of Artificial Intelligence, 115, 105151. DOI ↗ | Vinyals, O., Blundell, C., Lillicrap, T., Wierstra, D., & Kavukcuoglu, K. (2016). Matching Networks for One Shot Learning. Advances in Neural Information Processing Systems (NeurIPS), 29. link ↗ |
| Інші назви | transfer ensemble, multi-model transfer learning, ensemble of fine-tuned models, ETL | FSL, low-shot learning, k-shot learning, meta-learning for few examples |
| Пов'язані≠ | 6 | 4 |
| Підсумок≠ | Ensemble Transfer Learning combines multiple models that were each pre-trained on a large source domain and then fine-tuned on a target task. By aggregating the predictions of several independently fine-tuned models, it achieves higher accuracy and robustness than any single transferred model alone, especially when the target dataset is small. | Few-shot learning is a machine learning paradigm that trains models to recognize new classes or solve new tasks from only a handful of labeled examples — typically one to five — by leveraging prior knowledge acquired from a large, related training distribution. It is especially relevant in domains where labeling is expensive, scarce, or structurally limited. |
| ScholarGateНабір даних ↗ |
|
|