השוואת שיטות

	רשת קפסולות	זיקוק ידע (Knowledge Distillation)	מכונת וקטורים תומכים (סיווג)
תחום	למידה עמוקה	למידה עמוקה	למידת מכונה
משפחה	Machine learning	Machine learning	Machine learning
שנת המקור	2017	2015	1995
הוגה השיטה	Sabour, S., Frosst, N. & Hinton, G. E.	Hinton, G., Vinyals, O. & Dean, J.	Cortes, C. & Vapnik, V.
סוג	Deep learning architecture (vector capsules with dynamic routing)	Neural network compression (teacher–student)	Maximum-margin classifier (kernel method)
מקור מכונן	Sabour, S., Frosst, N. & Hinton, G. E. (2017). Dynamic Routing Between Capsules. Advances in Neural Information Processing Systems (NeurIPS).	Hinton, G., Vinyals, O. & Dean, J. (2015). Distilling the Knowledge in a Neural Network. NeurIPS Deep Learning Workshop.	Cortes, C. & Vapnik, V. (1995). Support-Vector Networks. Machine Learning, 20, 273–297.
כינויים	Kapsül Ağı (CapsNet), CapsNet, capsule net, dynamic routing network	Bilgi Damıtma (Knowledge Distillation), bilgi damıtma, teacher-student distillation, model distillation	Destek Vektör Makinesi (SVM — Sınıflandırma), support-vector network, SVM classifier, maximum-margin classifier
קשורות	4	5	5
תקציר	A Capsule Network (CapsNet) is a deep learning architecture introduced by Sara Sabour, Nicholas Frosst and Geoffrey Hinton in 2017 that organises neurons as vectors (capsules) rather than scalar activations, so that spatial hierarchy and pose (orientation) information are encoded directly. It was proposed to overcome the fragility of convolutional networks to changes in viewpoint.	Knowledge Distillation is a model-compression technique, introduced by Geoffrey Hinton and colleagues in 2015, that trains a small student model using the soft-label outputs of a large teacher model. Distilled models such as DistilBERT and TinyBERT reach roughly 97% of the larger model's performance while running far faster.	The Support Vector Machine, introduced by Corinna Cortes and Vladimir Vapnik in 1995, is a classifier that finds the optimal separating hyperplane between classes in a high-dimensional space. It chooses the boundary that leaves the widest possible margin to the nearest training points, which makes its decisions robust on new data.