So sánh phương pháp

	LoRA và PEFT	Generative Adversarial Network	Rừng ngẫu nhiên	Transformer Thị giác
Lĩnh vực	Học sâu	Học sâu	Học máy	Học sâu
Họ	Machine learning	Machine learning	Machine learning	Machine learning
Năm ra đời	2022	2014	2001	2021
Người khởi xướng	Hu, E. J. et al.; Lester, B. et al.	Goodfellow, I. et al.	Breiman, L.	Dosovitskiy, A. et al.
Loại	Parameter-efficient fine-tuning of large pretrained models	Generative deep learning (adversarial two-network game)	Ensemble (bagging of decision trees)	Transformer architecture for images (self-attention over patches)
Công trình gốc	Hu, E. J. et al. (2022). LoRA: Low-Rank Adaptation of Large Language Models. ICLR.	Goodfellow, I. et al. (2014). Generative Adversarial Nets. NeurIPS.	Breiman, L. (2001). Random Forests. Machine Learning, 45, 5–32.	Dosovitskiy, A. et al. (2021). An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. ICLR.
Tên gọi khác	LoRA ve PEFT — Parametre Verimli İnce Ayar, Low-Rank Adaptation, parameter-efficient fine-tuning, prefix tuning	Üretici Çekişmeli Ağ (GAN), GAN, generative adversarial nets, adversarial network	Rastgele Orman (Random Forest), rastgele orman, random decision forest, bagged tree ensemble	Görsel Transformer (ViT), görsel transformer, ViT, patch transformer for images
Liên quan	5	4	4	5
Tóm tắt	LoRA (Low-Rank Adaptation), introduced by Hu et al. in 2022, and the broader family of parameter-efficient fine-tuning (PEFT) methods adapt large pretrained language models to new tasks by training only a small number of extra parameters instead of every weight in the model. This makes fine-tuning possible with far less GPU memory and compute while leaving the original model largely untouched.	A Generative Adversarial Network (GAN), introduced by Ian Goodfellow and colleagues in 2014, produces realistic synthetic data through the competition of two neural networks — a generator and a discriminator. It is widely used for image synthesis, data augmentation, and distribution estimation.	Random Forest is an ensemble learning method, introduced by Leo Breiman in 2001, that grows many decision trees on bootstrap samples of the data and combines their votes to produce strong classification and regression. By pooling many slightly different trees, it produces more accurate and more stable predictions than any single tree.	The Vision Transformer (ViT), introduced by Dosovitskiy and colleagues in 2021, splits an image into fixed-size patches, treats those patches as a sequence, and applies the Transformer self-attention mechanism to image classification. Given enough training data, it surpasses convolutional neural networks (CNNs).