Machine Learning
Notes for the Machine Learning unit.
📄️ Machine Learning Basics
What is Machine Learning?
📄️ Simple Linear Regression
Definition
📄️ Multiple Linear Regression
Definition
📄️ Practical 1
Use pandas to prepare and manipulate data. Use matplotlib to visualize data. Use statsmodel to train a linear regression model, and improve the model by analyzing its performance.
📄️ Overfitting
📄️ Bias & Variance
📄️ Regularization
📄️ Neural Network Components
This page covers the building blocks of neural network and convolutional neural network.
📄️ ResNets
References
📄️ Fine Tuning
Deep networks need a lot of data to train. What can you do when you don't have much?
📄️ Data Augmentation
Deep networks need lots of data. It's one of the more annoying things about them. Collecting data is very painful, and is one of the more annoying things about machine learning. Data augmentation is a partial solution to both these annoyances.
📄️ Dimension Reduction
📄️ Principal Component Analysis
References
📄️ Linear Discriminant Analysis
📄️ t-SNE
📄️ Siamese Networks
📄️ Contrastive Loss
📄️ Triplet Loss
📄️ Embedding Size
📄️ K-Means
notes about K-Means clustering.
📄️ Gaussian Mixture Models
A Gaussian mixture model assumes that each cluster has its own normal (or Gaussian) distribution with parameters 𝜇𝑐 and 𝜎𝑐.
📄️ Selection of K
How do we select the number of clusters for K-Means and GMMs?
📄️ HAC
Hierarchical Agglomerative Clustering
📄️ DBScan
Density-Based Spatial Clustering of Applications with Noise.
📄️ Evaluating Clustering Performance
Use Purity, Completeness, and V-Measure to evaluate clustering performance.
📄️ Diarisation
Group identities in media
📄️ Auto Encoders
Encoder-Decoder
📄️ Multi-Task Learning
Multiple outputs from a deep neural network.
📄️ Semi-Supervised Learning
When you cannot label all the data.
📄️ Variational Auto-Encoders
Learn distributions
📄️ Assignment 1A
- Name: Baorong Huang
📄️ Assignment 1B
48 hour extension
📄️ Assignment 1C
Problem 1. Clustering and Recommendations. Problem 2. Multi-Task Learning