11_Machine Learning

Question 1

AI vs ML vs DL

Question 2

Machine Learning Options

Question 3

What is Machine Learning

• Process of combining inputs to produce useful predictions on never-before-seen data.
• Makes a machine learn from data to make predictions on future data, instead of programming every scenario.
• How it works:
  
  • Train a model with examples
  • Example = input + label
  • Training = adjust model to learn relationship between features and label - minimize error:
    
    • Optimize weights and biases (parameters) to different input features.
  • Feature = input variable(s)
  • Inference = apply trained model to unlabeled examples.
  • Separate test and training data ensures model is generalized for additional data.
    
    • Otherwise, leads to overfitting (only models to training data, not new data)

Question 4

Machine Learning Pipeline

Question 5

Features and Labels

Question 6

Curve Fitting

Question 7

Optimization using Gradient Descent

Question 8

Machine Learning Types

• Supervised learning
  
  • Apply labels to data (“cat”, “spam”)
  • Regression - Continuous, numeric variables:
    
    • Predict stock price, student test scores
  • Classification - categorical variables:
    
    • yes/no, decision tree
    • “is this email spam?” “is this picture a cat?”
  • Same types for dataset columns:
    
    • continuous (regression) and categorical (classification)
    • income, birth year = continuous
    • gender, country = categorical
• Unsupervised learning
  
  • Clustering - finding patterns
  • Not labeled or categorized
  • “Given the location of a purchase, what is the likely amount purchased?”
  • Heavily tied to statistics
• Reinforcement learning
  
  • Use positive/negative reinforcement to complete a task
    
    • Complete a maze, learn chess

Question 9

Supervised Learning

Question 10

Reinforcement Learning

Question 11

Model Type - Regression

Question 12

Model Type - Classification

Question 13

Model Type - Clustering

Question 14

Transfer Learning

Question 15

Overfitting

• Training model overfitted to training data: Unable to generalize with new data
• Training model fails to generalize: Accounting for slightly different but close enough data.
• Causes of Overfitting:
  
  • Not enough training data
    
    • Need more variety of samples
  • Too many features
    
    • Too complex
  • Model fitted to unnecessary features unique to training data, a.k.a “Noise”
• Solving for Overfitting:
  
  • Use more data:
    
    • Add more training data
    • More varied data allows for better generalization
  • Make the model less complex:
    
    • Use less (but more relevant) features = Feature Selection
    • Combine multiple co-dependant/redundant features into a single representative feature
      
      • This also helps reduce model training time
  • Remove noise
    
    • Increase regularization parameters
  • Regularization
  • Early Stopping
  • Cross Validation
  • Dropout Methods
• If data is scarce:
  
  • Use independent test data
  • Cross Validation

Question 16

Regularization

• Training: minimise (loss (data | model)
• Regularization: complexity (model)
  
  • L2 Regularization term
  • L1 Regularization term
• Training with Regularization: minimise (loss (data | model) + Lambda * complexity (model)
• Adds a penalty to a model as it becomes more complex
• Penalizing parameters = better generalization
• Cuts out noise and unimportant data, to avoid overfitting

Regularization Types

L1 and L2 regularization - Different approaches to tuning out noise. Each has different use case and purpose

• L1 - Lasso Regression: Assigns greater importance to more influential features
  
  • Shrinks less important features influence to zero
  • Good for models with many features, some more important than others
  • Example: Choosing features to predict likelihood of home selling:
    
    • House price more influential feature than carpet color
• L2 - Ridge Regression: Performs better when all the input features influence the output, and with all weights being roughly equal size.

Question 17

Hyperparameters

• Selection: Hyperparameter values needs to be specified before training begins
• Types of Hyperparameters:
  
  • Model hyperparameters relate directly to the model that is selected.
  • Algorithm hyperparameters relate to the training of the model.
• Training and Tuning: The process of finding the optimal, or near optimal values for hyperparameters.
• Not related to training data!
• Examples:
  
  • Batch size
  • Training epochs
  • Number of hidden layers in neural network
  • Number of nodes in hidden layers in neural network
  • Regularization type
  • Regularization rate
  • Learning rate aka steps size

Question 18

Feature Engineering

Transform data so it is fit for Machine Learning.

• Imputation (for missing data)
• Outliers and Feature Clipping
• One-hot Encoding (for categorical data)
• Linear Scaling
• Log Scaling
• Bucketing/Bucketization
• Feature prioritization

Question 19

Techniques Glossary

• Precision: formula to check how accurate the model is when most of the output are positives.
• Recall: formula to check how accurate the model is when most of the output are negatives.
• Gradient Descent: optimization algorithm to find the minimal value of a function. Gradient descent is used to find the minimal RMSE or cost function
• Dropout Regularization: regularization method to remove random selection of a fixed number of units in a neural network layer. More units dropped out, the stronger the regularization.