Usporedite metode

	Polu-nadzirana klasifikacija temeljena na RoBERTa modelu	Klasifikacija utemeljena na fino ugađenom RoBERTa-u
Područje	Duboko učenje	Duboko učenje
Obitelj	Machine learning	Machine learning
Godina nastanka	2019–2020	2019
Tvorac	Liu et al. (RoBERTa, 2019); semi-supervised adaptation by the NLP community	Liu, Y. et al. (Meta AI / University of Washington)
Vrsta	Semi-supervised fine-tuning of a pretrained language model	Pretrained transformer fine-tuned for classification
Temeljni izvor	Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019). RoBERTa: A Robustly Optimized BERT Pretraining Approach. arXiv preprint arXiv:1907.11692.	Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019). RoBERTa: A Robustly Optimized BERT Pretraining Approach. arXiv:1907.11692.
Drugi nazivi	Semi-supervised RoBERTa, RoBERTa with semi-supervised learning, SSL-RoBERTa classification, RoBERTa pseudo-label classification	RoBERTa fine-tuning, RoBERTa classifier, fine-tuned RoBERTa, RoBERTa sequence classification
Srodne	6	5
Sažetak	Semi-supervised RoBERTa-based classification combines a large pretrained RoBERTa language model with both a small labeled dataset and a larger pool of unlabeled text. By generating pseudo-labels or enforcing consistency on unlabeled examples, the method extracts supervisory signal from unannotated data, yielding stronger classifiers when ground-truth annotations are scarce.	Fine-tuned RoBERTa-based classification adapts the RoBERTa pretrained transformer — itself a robustly retrained variant of BERT — to a specific text classification task by appending a classification head and continuing training on labeled examples. It consistently achieves state-of-the-art or near-state-of-the-art performance on sentiment analysis, topic classification, toxicity detection, and similar NLP tasks.