roboticcam/machine-learning-notes

Features

• Machine Learning Education - Serves as a comprehensive educational resource for the mathematical and theoretical foundations of machine learning.
• Deep Learning Research - Explores advanced neural network architectures and generative models through a research-oriented lens.
• Machine Learning Foundations - Explores the foundational mathematics, probability, and statistics that form the theoretical basis of machine learning.
• Machine Learning Guides - Collects technical notes and code implementations covering the mathematics and architectures of machine learning.
• Probabilistic Graphical Models - Provides detailed technical documentation and implementations for Bayesian inference, Monte Carlo methods, and state space models.
• Markdown-Based Knowledge Bases - Stores theoretical explanations and mathematical formulas in version-controlled markdown files for portability.
• Annotated Code Implementations - Pairs theoretical mathematical explanations directly with executable code snippets to demonstrate model mechanics.
• Deep Learning Fundamentals - Covers core neural network concepts, including convolutional neural networks and loss functions, with practical implementation.
• Machine Learning Mathematics - Provides foundational mathematics covering model evaluation, decision trees, probability, and regression for machine learning.
• AI & Machine Learning Education - Provides foundational educational content covering artificial intelligence and machine learning algorithms and their practical implementations.
• Deep Learning Reference Implementations - Provides detailed technical explanations of neural networks paired with concrete code implementations.
• Deep Learning Architectures - Reviews technical specifications and structural compositions of convolutional and graph neural networks.
• Generative AI Learning Resources - Offers instructional notes and specialized research on generative models, transformers, and variational autoencoders.
• Generative AI Architectures - Examines transformer structures and attention mechanisms by implementing key-value caching and core model mechanics.
• Study Guides - Provides instructional material on 3D reconstruction and epipolar geometry for depth estimation.
• Deep Learning Optimization - Evaluates research on implicit bias and duality to improve convergence for stochastic gradient descent.
• Gradient Descent Algorithms - Analyzes optimization algorithms including Lagrangian duality, KKT conditions, and conjugate gradient descent.
• Probabilistic Models - Models complex data using expectation maximization, Markov Chain Monte Carlo, and variational inference.
• Transformer Architecture Implementation - Implements sequence-to-sequence architectures including rotary positional embeddings and multi-head latent attention.
• Bayesian Inference - Estimates unknown variables by applying Monte Carlo methods, particle filtering, and Bayesian non-parametrics.
• Geometry and Vision - Provides resources for reconstructing 3D spaces using camera models and epipolar geometry.
• Curricula - Provides instructional material on 3D reconstruction, camera models, and epipolar geometry.
• Learning Paths - Structures theoretical learning paths starting from foundational linear algebra and probability before moving to complex architectures.
• Mathematics Study Guides - Provides lecture notes and formula derivations for intermediate ML math, including expectation maximization and variational inference.
• Natural Language Processing Resources - Provides educational content and academic breakdowns of natural language processing, specifically focusing on word embeddings and attention mechanisms.
• Reinforcement Learning Theory - Explores the mathematical and theoretical foundations of reinforcement learning, including Markov Decision Processes and policy gradients.
• Curriculum Decomposition - Breaks complex subjects like 3D vision and probabilistic inference into discrete notes for incremental learning.
• Topic-Based Resource Organization - Organizes educational content into a nested folder structure based on machine learning domains and mathematical prerequisites.
• Epipolar Geometry - Implements depth and pose estimation using camera models and epipolar geometry.
• Monte Carlo Sampling - Executes probabilistic simulations through inverse CDF sampling, importance sampling, and particle filters.
• Technical Knowledge Maps - Connects diverse fields like 3D vision and probabilistic inference through a unified system of interlinked technical notes.

Star history