Byte-Sized Deep Learning Series: Demystifying Neural Networks
- Created By shambhvi
- Posted on May 1st, 2025
- Overview
- Prerequisites
- Audience
- Curriculum
Description:
Step into the world of deep learning by building your first Artificial Neural Network!
In this 90-minute, beginner-friendly session, you’ll uncover how neural networks learn, why activation functions are critical, and how to build a real model using TensorFlow and Keras.
If you know machine learning basics but want to start constructing real deep learning models, this is the perfect launchpad.
Duration: 90 mins
Course Code: BDT493
Learning Objectives:
After this course, you will be able to:
- Demystifying Artificial Neural Networks (ANNs)
- How do ANNs learn?
- Building a basic ANN
Must have some python programming experience and NumPy. Students must be comfortable with TensorFlow basics
Machine learning enthusiasts and practitioners who are familiar with basic ML concepts (e.g., models, datasets, training loops). Students new to deep learning and neural networks, wanting a clear and intuitive understanding of how networks learn
Course Outline:
- Demystifying Artificial Neural Networks
- What is a neural network? How does it mimic the brain?
- Core components of neural network: neurons, layers, weights, biases
- Understanding Forward and Backward pass
- How do networks learn? Loss and Optimization
- The learning goal: minimizing the loss function
- Loss functions: MSE, Cross Entropy
- Optimizers: Adam, SGD
- Hands-on: Creating and Visualizing Loss
- Building first neural network with Keras
- A brief introduction to Keras high-level API
- Build a simple neural network
- Compiling the model (specifying Loss and Optimizer)
- Hands-on: Build a simple neural network
Training material provided: Yes (Digital format)
Hands-on Lab: Instructions will be provided to install Jupyter notebook and other required python libraries. Students can opt to use ‘Google Colaboratory’ if they do not want to install these tools