Getting Started
| Term | Definition |
|---|---|
| Agent | An entity that perceives its environment and takes actions to achieve goals |
| Percept | What an agent senses at a given moment from the environment |
| Heuristic | A rule-of-thumb or informed guess that helps solve problems efficiently |
| Overfitting | Model learns training data too well, including noise, and fails on new data |
| Underfitting | Model is too simple to capture patterns in the data |
| Bias | Systematic error in predictions; also refers to societal bias in AI systems |
| Variance | Sensitivity of model to fluctuations in training data |
| Feature | An individual measurable property of a data point used for learning |
| Label | The target output in supervised learning (e.g., "cat" or "dog") |
| Training | The process of teaching a model using labeled or unlabeled data |
| Inference | Using a trained model to make predictions on new data |
| Latent Space | Compressed representation of data learned by a neural network |
| Token | A unit of text processed by an LLM (word, subword, or character) |
| Attention | Mechanism that lets models focus on relevant parts of input when generating output |
| Transfer Learning | Using knowledge from one task to improve learning on a related task |
| Fine-tuning | Adapting a pre-trained model for a specific downstream task |
| Reinforcement Learning | Learning through trial-and-error with rewards and penalties |
| Hallucination | When an AI generates plausible but incorrect information |