방법 비교
선택한 방법을 나란히 검토하세요. 서로 다른 행은 강조 표시됩니다.
| 약한 지도 학습 텍스트 요약× | 자기 지도 학습× | |
|---|---|---|
| 분야≠ | 딥러닝 | 머신러닝 |
| 계열 | Machine learning | Machine learning |
| 기원 연도≠ | 2015–2020 | 2018–2020 |
| 창시자≠ | Multiple independent research groups (NLP community, 2010s–2020s) | LeCun, Y. and community (formalized ~2018–2020) |
| 유형≠ | Semi-supervised / weakly supervised NLP training paradigm | Representation learning paradigm |
| 원전≠ | Amplayo, R. K., & Lapata, M. (2020). Unsupervised Opinion Summarization with Noisy Autoencoder. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 1934–1945. link ↗ | LeCun, Y. & Misra, I. (2022). Self-supervised learning: The dark matter of intelligence. Meta AI Blog. https://ai.facebook.com/blog/self-supervised-learning-the-dark-matter-of-intelligence/ link ↗ |
| 별칭 | weak supervision summarization, distantly supervised summarization, noisy-label summarization, pseudo-label summarization | SSL, self-supervised pre-training, pretext-task learning, unsupervised representation learning |
| 관련≠ | 1 | 3 |
| 요약≠ | Weakly supervised text summarization trains abstractive or extractive summarization models without manually annotated reference summaries. Instead of costly human labels, it exploits weak signals — heuristic rules, distant supervision, noisy automatic labels, or self-supervised objectives — to guide sequence-to-sequence or transformer models toward producing coherent, concise summaries of input documents. | Self-supervised learning (SSL) is a machine-learning paradigm that generates its own supervisory signal directly from unlabeled data by defining an auxiliary pretext task — such as predicting masked words, rotating images, or contrasting augmented views — and uses the learned representations as a powerful starting point for downstream tasks with minimal labeled examples. |
| ScholarGate데이터셋 ↗ |
|
|