方法对比

	数据增强 (Data Augmentation)	对抗训练	迁移学习
领域	深度学习	深度学习	机器学习
方法族	Machine learning	Machine learning	Machine learning
起源年份	2019	2018	2010 (formalized); 1990s (early roots)
提出者	Connor Shorten & Taghi Khoshgoftaar	Aleksander Madry et al.	Pan, S. J. & Yang, Q. (survey); Bengio, Y. (deep learning framing)
类型	Regularization / data preprocessing technique	Robust optimization training procedure	Learning paradigm
开创性文献	Shorten, C., & Khoshgoftaar, T. M. (2019). A survey on image data augmentation for deep learning. Journal of Big Data, 6, 60.	Madry, A., Makelov, A., Schmidt, L., Tsipras, D., & Vladu, A. (2018). Towards deep learning models resistant to adversarial attacks. International Conference on Learning Representations (ICLR).	Pan, S. J., & Yang, Q. (2010). A Survey on Transfer Learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359.
别名	Training Data Augmentation, Image Augmentation, Veri Artırma, Synthetic Data Augmentation	Min-Max Robust Training, PGD Adversarial Training, Robust Empirical Risk Minimization, Hasımsal Eğitim	TL, domain adaptation, fine-tuning, pre-trained model adaptation
相关	2	3	3
摘要	Data augmentation is a family of techniques that artificially expands a training dataset by applying label-preserving transformations to existing samples. Originally systematized for image classification tasks, it is now applied broadly across vision, text, audio, and tabular domains. It emerged as a practical answer to the chronic scarcity of labeled data in supervised deep learning and remains a standard preprocessing step in modern neural network pipelines.	Adversarial Training is a robust optimization procedure for deep neural networks in which the model is trained not on clean data alone but on worst-case perturbed inputs crafted during training. Formalized by Madry et al. (2018) as a min-max saddle-point problem, the method uses Projected Gradient Descent (PGD) to generate strong adversarial examples within a bounded Lp perturbation set before each gradient update, forcing the network to learn decision boundaries that are stable under such perturbations.	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.