השוואת שיטות

	למידת מטריקה מונחית-למחצה	Transfer Learning
תחום	למידת מכונה	למידת מכונה
משפחה	Machine learning	Machine learning
שנת המקור	2007–2008	2010 (formalized); 1990s (early roots)
הוגה השיטה	Yeung, D.-Y. & Chang, H.; Davis, J. V. & Dhillon, I. S.	Pan, S. J. & Yang, Q. (survey); Bengio, Y. (deep learning framing)
סוג	Hybrid supervised/unsupervised distance learning	Learning paradigm
מקור מכונן	Yeung, D.-Y., & Chang, H. (2007). A kernel approach for semi-supervised metric learning. IEEE Transactions on Neural Networks, 18(1), 141–149.	Pan, S. J., & Yang, Q. (2010). A Survey on Transfer Learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359.
כינויים	SSML, semi-supervised distance learning, constrained metric learning, weakly supervised metric learning	TL, domain adaptation, fine-tuning, pre-trained model adaptation
קשורות	5	3
תקציר	Semi-supervised metric learning learns a task-adapted distance function by combining a small set of labeled pairwise constraints — must-link and cannot-link pairs — with the geometric structure of a much larger pool of unlabeled data. The result is a Mahalanobis-style or kernel-based distance that reflects both supervision and data topology, improving downstream tasks such as nearest-neighbor classification and clustering.	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.