X

A family of machine learning approaches that train models to recognize or perform tasks with very few (few-shot), one (one-shot), or zero (zero-shot) labeled examples. X-shot learning is critical for real-world AI deployment where labeled training data is scarce or expensive to collect. It is the foundation of modern LLM capabilities — a GPT model can perform a new task from just a few examples given in the prompt, without any retraining.

An extremely efficient and widely-used machine learning algorithm based on gradient boosting decision trees. XGBoost consistently wins data science competitions and is a go-to tool for structured/tabular data prediction tasks — such as credit scoring, property valuation, churn prediction, and fraud detection. Unlike neural networks, XGBoost models are fast to train, interpretable, and perform well on smaller datasets.