Beginner Friendly Intro to AI Course (Part 2)

Explore machine learning basics, including supervised, unsupervised, and reinforcement learning, via a linear regression example predicting grades from hours studied, with training/testing splits, mean squared error, and scikit-learn.

Learn logistic regression as a classification method using a logistic function to predict pass or fail from hours studied. Split into training and testing, then apply a threshold to decide.

Explore regression methods, including linear and logistic regression, using train-test splits and accuracy score, then compare decision trees with recursive binary splits for complex boundaries.

Explore overfitting in decision trees and how max depth and random forests prevent it, implement classifiers with scikit-learn, and introduce unsupervised learning with clustering and dimensionality reduction, including k-means.

Explore evaluation methods for machine learning, focusing on classification metrics like confusion matrix, accuracy, precision, recall, and F1 score, and visualize performance with ROC curves and AUC.

Explore regression metrics beyond mean squared error, including r squared, and watch a five-minute video explaining how R squared works.

Explore core computer vision concepts like object recognition, image analysis, AR/VR, robotics, and surveillance. Understand convolutional neural networks and neural networks, transfer learning, and training with gradient descent.

Explore convolutional neural networks (CNNs) for computer vision, using input, hidden, and output layers with kernels to detect edges, corners, and textures. Use ReLU and pooling to learn features efficiently.

Explore activation functions in CNNs, including ReLU, and learn how different types of activation functions enhance machine learning performance.

Move from activation functions to optimizers as SGD, Adam, RMSprop, and Adagrad update neural network parameters to minimize the loss function, using adaptive learning rates, momentum, and transfer learning.

Explore transfer learning with pre-trained models for image classification, object detection, and NLP tasks, including vgg, resnet, inception/googLeNet, mobilenet, bert, gpt, yolo, densenet, and exception.