Neural Networks aren't as complicated as they seem to be. Anyone who can write a few lines of Python code can easily whip up an elegant model to predict text or recognise speech using Neural Networks with the help of the Keras library. I will go through how any beginner can use Keras to start their journey as Neural Network pros.

Attendees of this talk will develop an understanding of how neural networks work and how they can implement these complex models with simplicity using the Keras library in Python.

This presentation will go over how Neural Networks work. I will talk about the differences between RNNs, LSTMs etc and help you learn when to use which type of network. After a brief theoretical introduction to the domain, I will move on to actually implementing Neural Networks in Python using Keras. I will take it right from setting up Keras on your machine all the way to analysing the accuracy of your model. You will learn about the different backends that Keras works with. You will be exposed to implementing the different models Keras has to offer. I will talk about Optimizer functions like Stochastic Gradient Descent, RMS Prop etc. You will learn about loss functions like Cross Entropy loss. You will see how to modify your Keras model on the fly based on the results of the Neural Network (how accurate it is). We will do a demo of a basic neural network on a sample data set where attendees of this talk will get hands on experience with fine tuning the parameters of a neural network to optimize results. With the help of Keras you will be able to jump right into developing complex models using Neural Networks without the hassle of all the boiler plate code you would need without a handy library like Keras.