A Novel Context-Aware Framework for Intelligent and Automated Data Augmentation in Machine Learning
Abstract: Automated data augmentation has become an important technique for improving the performance and generalization of machine learning models by creating diverse training samples. This article examines major approaches to automated data augmentation, including analytical transformations, deep learning methods, generative models, and optimization techniques within Automated Machine Learning (AutoML) frameworks. It discusses how reinforcement learning, Bayesian optimization, and evolutionary algorithms help identify suitable augmentation strategies. Particular attention is given to generative artificial intelligence, including large language models and diffusion models, which can produce synthetic data for computer vision, natural language processing, and other applications. Dataset distillation is also examined as an alternative approach for reducing training data while preserving useful information. Comparative findings from established image classification datasets show that automated augmentation methods generally perform better than traditional manual techniques. However, high computational requirements, complex optimization processes, and the risk of altering the original meaning of data remain major challenges. The discussion highlights the need for efficient, reliable, and context-sensitive augmentation methods that can support machine learning applications across different domains.
Keywords: Automated Data Augmentation, Automated Machine Learning, Generative Artificial Intelligence, Synthetic Data Generation, Dataset Distillation
How to Cite?: Putty Srividya, Dr N. Subramanyam, "A Novel Context-Aware Framework for Intelligent and Automated Data Augmentation in Machine Learning", Volume 15 Issue 1, January 2026, International Journal of Science and Research (IJSR), Pages: 1854-1861, https://www.ijsr.net/getabstract.php?paperid=SR260112105952, DOI: https://dx.doi.org/10.21275/SR260112105952