Ground Truth (machine learning)

Ground truth in machine learning refers to the real-world data or accurate information used as a benchmark to train and evaluate models. It’s the correct, verified data that the model aims to learn from or predict, such as labeled images, correct diagnoses, or actual measurements. By comparing a model’s predictions to ground truth, developers can assess how well the AI is performing and make improvements. Essentially, ground truth serves as the reference point for training and validating machine learning systems, ensuring their outputs are aligned with reality.