Comparar métodos

	Aprendizado Autossupervisionado	Aprendizado com Poucos Exemplos	Aprendizado Semi-supervisionado	Aprendizagem por Transferência
Área	Aprendizado de máquina	Aprendizado de máquina	Aprendizado de máquina	Aprendizado de máquina
Família	Machine learning	Machine learning	Machine learning	Machine learning
Ano de origem	2018–2020	2011–2017	1970s–2006 (formalized)	2010 (formalized); 1990s (early roots)
Autor original	LeCun, Y. and community (formalized ~2018–2020)	Lake, B. M.; Vinyals, O.; Finn, C. et al.	Vapnik, V. N. and others (community of researchers, 1970s–2000s)	Pan, S. J. & Yang, Q. (survey); Bengio, Y. (deep learning framing)
Tipo	Representation learning paradigm	Meta-learning / low-data learning paradigm	Learning paradigm	Learning paradigm
Fonte seminal	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/	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.	Chapelle, O., Scholkopf, B., & Zien, A. (Eds.) (2006). Semi-Supervised Learning. MIT Press. ISBN: 978-0-262-03358-9	Pan, S. J., & Yang, Q. (2010). A Survey on Transfer Learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359.
Outros nomes	SSL, self-supervised pre-training, pretext-task learning, unsupervised representation learning	FSL, low-shot learning, k-shot learning, meta-learning for few examples	SSL, semi-supervised machine learning, transductive learning, label-efficient learning	TL, domain adaptation, fine-tuning, pre-trained model adaptation
Relacionados	3	4	5	3
Resumo	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.	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.	Semi-supervised learning (SSL) is a machine learning paradigm that trains models using a small set of labeled examples together with a much larger pool of unlabeled data. By leveraging the structure inherent in unlabeled data, SSL achieves accuracy closer to fully supervised models while requiring far fewer costly manual labels — making it practical when labeling is expensive, slow, or resource-constrained.	Transfer learning is a machine learning paradigm in which knowledge gained from training a model on a source task or domain is reused to improve learning on a different but related target task or domain. It is especially powerful when labeled data for the target task is scarce, and it underlies most modern deep learning applications in computer vision, natural language processing, and beyond.