Comparar métodos

	Vision Transformer com Supervisão Fraca	Destilação de Conhecimento
Área	Aprendizado profundo	Aprendizado profundo
Família	Machine learning	Machine learning
Ano de origem	2021–2022	2015
Autor original	Dosovitskiy et al. (ViT); weak supervision paradigm from Zhou and others	Hinton, G., Vinyals, O. & Dean, J.
Tipo	Self-attention image model with weakly supervised training	Neural network compression (teacher–student)
Fonte seminal	Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., & Houlsby, N. (2021). An image is worth 16x16 words: Transformers for image recognition at scale. In International Conference on Learning Representations (ICLR).	Hinton, G., Vinyals, O. & Dean, J. (2015). Distilling the Knowledge in a Neural Network. NeurIPS Deep Learning Workshop.
Outros nomes	WS-ViT, weakly supervised ViT, weak supervision with vision transformer, ViT with weak labels	Bilgi Damıtma (Knowledge Distillation), bilgi damıtma, teacher-student distillation, model distillation
Relacionados	4	5
Resumo	Weakly Supervised Vision Transformer (WS-ViT) trains a Vision Transformer on image data that lacks precise pixel-level annotations, instead using cheaper, noisier supervision such as image-level class tags, bounding boxes, or web-scraped text. The global self-attention mechanism of the transformer makes it especially capable of localising objects and learning discriminative features from these incomplete labels.	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.