Сравнение на методи
Прегледайте избраните методи един до друг; редовете с разлики са откроени.
| Рекурентна невронна мрежа с адаптация към домейн× | Адаптивно към домейна класифициране, базирано на BERT× | |
|---|---|---|
| Област | Дълбоко обучение | Дълбоко обучение |
| Семейство | Machine learning | Machine learning |
| Година на възникване≠ | 2010s | 2019–2020 |
| Създател≠ | Ganin et al.; Pan & Yang (domain adaptation frameworks applied to RNNs) | Gururangan et al. (2020); earlier domain-specific instances include Lee et al. (2020) — BioBERT |
| Тип≠ | Domain-adaptive sequential model | Domain-adaptive pre-training followed by supervised fine-tuning |
| Основополагащ източник≠ | Ganin, Y., Ustunova, E., Ajakan, H., Germain, P., Larochelle, H., Laviolette, F., Marchand, M., & Lempitsky, V. (2016). Domain-adversarial training of neural networks. Journal of Machine Learning Research, 17(59), 1–35. link ↗ | Gururangan, S., Marasovic, A., Swayamdipta, S., Lo, K., Beltagy, I., Downey, D., & Smith, N. A. (2020). Don't Stop Pretraining: Adapt Language Models to Domains and Tasks. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL 2020), 8342–8360. DOI ↗ |
| Други названия | DA-RNN, domain-adaptive RNN, domain-adapted recurrent network, cross-domain RNN | DAPT BERT classification, domain-adaptive pre-training, domain-specific BERT fine-tuning, BERT DAPT |
| Свързани | 6 | 6 |
| Резюме≠ | A Domain-adaptive Recurrent Neural Network (DA-RNN) is a recurrent neural network trained on a source domain and adapted to a target domain using domain adaptation techniques such as adversarial training, feature alignment, or fine-tuning. It enables sequential models to generalise across domains when labeled target-domain data is scarce or unavailable. | Domain-adaptive BERT-based classification extends the standard fine-tuning pipeline by first continuing BERT's masked-language-model pre-training on a large corpus of in-domain unlabeled text, then fine-tuning the adapted model on labeled examples for the target classification task. This two-stage approach closes the vocabulary and distributional gap between BERT's general pre-training corpus and specialized domains such as biomedicine, law, finance, or social-media text. |
| ScholarGateНабор от данни ↗ |
|
|