At the sub group meeting on 20th May, we discussed the need for a glossary, so as a starter for ten Sam Williamson asked Chat GPT to generate one. This is what it came up with:

Here’s a glossary of commonly used AI terms, broken down into categories for clarity:
________________________________________

General AI Concepts
• Artificial Intelligence (AI): The simulation of human intelligence processes by machines, especially computer systems.
• Machine Learning (ML): A subset of AI that enables machines to learn from data and improve from experience without being explicitly programmed.
• Deep Learning: A type of machine learning that uses neural networks with many layers (deep neural networks) to analyze complex data.
• Neural Network: A system of algorithms modeled after the human brain, designed to recognize patterns.
• Supervised Learning: Machine learning where the model is trained on labeled data (inputs paired with correct outputs).
• Unsupervised Learning: The model learns patterns from data without labeled responses.
• Reinforcement Learning: A learning approach where agents take actions in an environment to maximize a reward signal.
________________________________________

AI Types and Systems
• Narrow AI (Weak AI): AI specialized in a single task (e.g., Siri, facial recognition).
• General AI (Strong AI): Hypothetical AI that can perform any intellectual task a human can do.
• AGI (Artificial General Intelligence): A theoretical form of AI with the ability to understand, learn, and apply knowledge in a generalized way.
• LLM (Large Language Model): A type of AI model trained on massive text datasets to understand and generate human-like language (e.g., GPT-4).
________________________________________

Model Training & Evaluation
• Training Data: The dataset used to teach a model during the learning process.
• Validation Data: Used to tune model parameters and avoid overfitting.
• Testing Data: Independent data used to assess the performance of the model after training.
• Overfitting: When a model learns the training data too well, including noise, reducing its ability to generalize.
• Underfitting: When a model is too simple to capture patterns in the data, leading to poor performance.
• Bias: Systematic error introduced by incorrect assumptions in the learning algorithm.
• Variance: Error due to sensitivity to small fluctuations in the training set.
________________________________________

Technical Components
• Algorithm: A set of rules or instructions a machine follows to solve a problem.
• Epoch: One complete pass through the entire training dataset.
• Gradient Descent: An optimization algorithm used to minimize the error of a model by adjusting parameters.
• Loss Function: A method of evaluating how well a model’s prediction matches the expected output.
• Backpropagation: A method for updating weights in a neural network based on error gradients.
________________________________________

Metrics & Evaluation
• Accuracy: The ratio of correct predictions to total predictions.
• Precision: The ratio of correctly predicted positive observations to total predicted positives.
• Recall: The ratio of correctly predicted positive observations to all actual positives.
• F1 Score: The harmonic mean of precision and recall.
• ROC Curve / AUC: Tools for visualizing and quantifying classification performance.
________________________________________

Advanced Concepts
• Transformer: A neural network architecture that uses self-attention mechanisms, foundational to modern LLMs like GPT.
• Attention Mechanism: Allows models to focus on relevant parts of the input when making predictions.
• Prompt Engineering: The practice of designing effective inputs to elicit desired responses from LLMs.
• Fine-Tuning: Adjusting a pre-trained model with new, often domain-specific, data.
________________________________________

Ethics & Philosophy
• Bias in AI: When an AI model reflects or amplifies social biases present in the data.
• Explainability (XAI): The ability to explain how and why an AI system makes decisions.
• Fairness: The goal of ensuring AI systems do not discriminate against individuals or groups.
• Transparency: How openly the workings and decisions of an AI model can be understood.
• AI Alignment: Ensuring AI systems' goals and behaviors align with human values.
________________________________________