방법 비교

	Direct Preference Optimization	Masked Autoencoders	QLoRA
분야	딥러닝	딥러닝	딥러닝
계열	Machine learning	Machine learning	Machine learning
기원 연도	2023	2021	2023
창시자	Rafael Rafailov	Kaiming He	Tim Dettmers
유형	Training methodology	Neural network architecture	Training methodology
원전	Rafailov, R., Sharma, A., Mitchell, E., Manning, C. D., Ermon, S., & Finn, C. (2023). Direct preference optimization: Your language model is secretly a reward model. arXiv preprint arXiv:2305.18290.	He, K., Chen, X., Xie, S., Li, Y., Dollár, P., & Girshick, R. (2022). Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 16000-16009).	Dettmers, T., Pagnoni, A., Holtzman, A., & Contrastive, L. (2023). QLoRA: Efficient finetuning of quantized LLMs. arXiv preprint arXiv:2305.14314.
별칭	DPO, Direct preference	MAE, Vision MAE	QLoRA, Quantized LoRA
관련	4	4	4
요약	Direct Preference Optimization (DPO) is a training method introduced by Rafailov et al. in 2023 that aligns language models with human preferences without requiring an explicit reward model. By directly optimizing for preference pairs (better response vs worse response), DPO simplifies the training pipeline compared to reinforcement learning from human feedback (RLHF).	Masked Autoencoders (MAE) is a self-supervised learning approach introduced by He et al. in 2021 that masks random patches of an image and trains a model to reconstruct the missing content. Adapting the masked language modeling paradigm from NLP to vision, MAE learns rich visual representations by solving a challenging reconstruction task without requiring labels.	QLoRA is an efficient fine-tuning method introduced by Dettmers et al. in 2023 that enables fine-tuning large language models using quantization and low-rank adaptation. By combining 4-bit quantization with LoRA, QLoRA reduces memory requirements by 75%, enabling fine-tuning of 65B-parameter models on single GPUs.