Supervised Learning
A type of machine learning where the model learns from labeled training data to make predictions or classify data points.
Underfitting
Occurs when a model is too simple to capture the underlying patterns in data, resulting in poor performance on both training and test sets.
Overfitting
When a model learns the training data too well, including noise and outliers, and performs poorly on new, unseen data.
Gradient Descent
An optimization algorithm used to minimize the loss function by iteratively updating model parameters in the opposite direction of the gradient.
Evaluation Metrics
Methods to assess model performance such as accuracy, precision, recall, F1-score, and confusion matrix.
Activation Function
Functions like ReLU, Sigmoid, or Tanh that introduce non-linearity into neural networks, allowing them to learn complex patterns.
CNN
Convolutional Neural Networks are specialized deep learning models primarily used for processing grid-like data such as images.
Deep Learning
A subset of machine learning using neural networks with many layers to model complex patterns in large datasets.
Attention
A mechanism that enables models to focus on relevant parts of the input, crucial in NLP and transformer architectures.
Transformers
Deep learning models that rely entirely on attention mechanisms and are foundational to modern NLP (e.g., BERT, GPT).
Transfer Learning
A technique where a pre-trained model is reused on a new but related problem, speeding up training and improving performance.
Reinforcement Learning
A learning paradigm where agents take actions in an environment to maximize cumulative reward over time.