👉 Fairy engineering is a concept from the field of artificial intelligence, particularly in the context of generative models like GANs (Generative Adversarial Networks) and VAEs (Variational Autoencoders). It refers to the practice of using a small, carefully crafted set of "fairy godmothers" or "fairy engineers"—specific, expertly designed models—to guide the learning process of a larger, more complex model. These fairy engineers are typically human experts with deep domain knowledge who create high-quality, synthetic data samples that the primary model learns from. By providing these expertly crafted inputs, fairy engineering helps the main model to generate more accurate, diverse, and high-quality outputs, accelerating learning and improving performance, especially in scenarios where data is scarce or of poor quality.
