Coursera Bets $100 Million That Andrew Ng Can Do What Coursera Can’t

Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

CodeSignal

The MLP Architecture: Activations & Initialization

via CodeSignal

Go to class Write review

Overview

This course builds upon single layers to construct a complete Multi-Layer Perceptron (MLP). You'll learn to stack layers, explore different activation functions like ReLU and Softmax, and understand the importance of weight initialization for effective training.

Syllabus

  • Unit 1: Building a Multilayer Perceptron
    • Complete the MLP Forward Pass
    • Fix MLP Layer Dimension Mismatch
    • Building the MLP Class from Scratch
    • Adding a Fourth Layer to the MLP
    • Building a Complete Multi-Layer Perceptron from Scratch
  • Unit 2: ReLU Activation and Initialization
    • Fixing the ReLU Activation Function for Matrix Operations
    • Adding ReLU Activation Support to DenseLayer
    • Implementing the ReLU Activation Function
  • Unit 3: Output Layer Activations
    • Implementing Numerically Stable Softmax Activation
    • Validating Softmax Probability Distributions
    • Complete the Linear Activation Function
    • Fix the Softmax Activation Bug
    • Building Classification and Regression Networks with Output Activations
  • Unit 4: Weight Initialization Strategies
    • Implementing Random Scaled Weight Initialization
    • Fix He Uniform Weight Initialization
    • Implementing Xavier Normal Weight Initialization in C++
    • Implementing He Uniform Weight Initialization for ReLU Networks

Reviews

Start your review of The MLP Architecture: Activations & Initialization

Go to class

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Sign up for free
Someone learning on their laptop while sitting on the floor.