Porównaj metody
Przeglądaj wybrane metody obok siebie; wiersze, które się różnią, są wyróżnione.
| Destylacja wiedzy× | Longformer / BigBird× | |
|---|---|---|
| Dziedzina | Uczenie głębokie | Uczenie głębokie |
| Rodzina | Machine learning | Machine learning |
| Rok powstania≠ | 2015 | 2020 |
| Twórca≠ | Hinton, G., Vinyals, O. & Dean, J. | Beltagy, Peters & Cohan (Longformer); Zaheer et al. (BigBird) |
| Typ≠ | Neural network compression (teacher–student) | Sparse-attention Transformer for long sequences |
| Źródło pierwotne≠ | Hinton, G., Vinyals, O. & Dean, J. (2015). Distilling the Knowledge in a Neural Network. NeurIPS Deep Learning Workshop. link ↗ | Beltagy, I., Peters, M. E. & Cohan, A. (2020). Longformer: The Long-Document Transformer. arXiv. link ↗ |
| Inne nazwy | Bilgi Damıtma (Knowledge Distillation), bilgi damıtma, teacher-student distillation, model distillation | Uzun Dizi Transformer (Longformer / BigBird), uzun dizi transformer, long-document transformer, sparse-attention transformer |
| Pokrewne≠ | 5 | 4 |
| Podsumowanie≠ | Knowledge Distillation is a model-compression technique, introduced by Geoffrey Hinton and colleagues in 2015, that trains a small student model using the soft-label outputs of a large teacher model. Distilled models such as DistilBERT and TinyBERT reach roughly 97% of the larger model's performance while running far faster. | Long-sequence Transformers such as Longformer (Beltagy, Peters & Cohan, 2020) and BigBird (Zaheer et al., 2020) replace the standard Transformer's O(n²) attention with sparse attention patterns that scale linearly, O(n), with sequence length. This lets a single model attend over thousands of tokens — full documents, legal texts, or genomic sequences — that would not fit a conventional Transformer. |
| ScholarGateZbiór danych ↗ |
|
|