方法对比

	自监督少样本学习	Siamese Neural Network	迁移学习
领域	机器学习	深度学习	机器学习
方法族	Machine learning	Machine learning	Machine learning
起源年份	2019	1993	2010 (formalized); 1990s (early roots)
提出者	Gidaris, S. et al.; Su, J.-C. et al. (concurrent seminal works)	Jane Bromley & Yann LeCun et al.; popularized by Koch et al.	Pan, S. J. & Yang, Q. (survey); Bengio, Y. (deep learning framing)
类型	Hybrid learning paradigm (self-supervised pretraining + few-shot adaptation)	Deep metric-learning architecture	Learning paradigm
开创性文献	Gidaris, S., Bursuc, A., Komodakis, N., Perez, P., & Cord, M. (2019). Boosting Few-Shot Visual Learning with Self-Supervision. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 8059–8068.	Bromley, J., Guyon, I., LeCun, Y., Säckinger, E., & Shah, R. (1993). Signature verification using a 'Siamese' time delay neural network. Advances in Neural Information Processing Systems, 6.	Pan, S. J., & Yang, Q. (2010). A Survey on Transfer Learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359.
别名	SSL-FSL, self-supervised meta-learning, unsupervised few-shot learning, self-supervised prototypical learning	twin network, Siamese neural network, contrastive metric network, Siyam ağı	TL, domain adaptation, fine-tuning, pre-trained model adaptation
相关	2	1	3
摘要	Self-supervised Few-shot Learning (SSL-FSL) combines self-supervised pretraining on large unlabeled corpora with few-shot meta-learning so that a model can recognize new categories from only a handful of labeled examples. By learning rich, transferable representations without expensive annotation, SSL-FSL addresses the fundamental bottleneck of supervised few-shot methods: the need for labeled support data at scale.	A Siamese network is a deep architecture with two (or more) identical, weight-sharing branches that map inputs into an embedding space where similar inputs land close together and dissimilar ones far apart. Introduced by Bromley, LeCun, and colleagues in 1993 for signature verification and revived by Koch et al. (2015) for one-shot image recognition, it learns a similarity metric rather than fixed class labels, making it ideal for verification, matching, and few-shot tasks.	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.