INTUITIVE AUDIO DEEP LEARNING SERIES

A Gentle Guide to enhancing Spectrogram features for optimal performance. Also Data Augmentation, in Plain English

Image for post
Image for post
Photo by Vidar Nordli-Mathisen on Unsplash

This is the third article in my series on audio deep learning. So far we’ve learned how sound is represented digitally, and that deep learning architectures usually use a spectrogram of the sound. We’ve also seen how to pre-process audio data in Python to generate Mel Spectrograms.

In this article, we will take that a step further and enhance our Mel Spectrogram by tuning its hyper-parameters. We will also look at Augmentation techniques for audio data. Both of these are essential aspects of data preparation in order to get better performance from our audio deep learning models.

Here’s a quick…

INTUITIVE AUDIO DEEP LEARNING SERIES

A Gentle Guide to processing audio in Python. What are Mel Spectrograms and how to generate them, in Plain English.

Image for post
Image for post
Photo by Jordan on Unsplash

This is the second article in my series on audio deep learning. Now that we know how sound is represented digitally, and that we need to convert it into a spectrogram for use in deep learning architectures, let us understand in more detail how that is done and how we can tune that conversion to get better performance.

Since data preparation is so critical, particularly in the case of audio deep learning models, that will be the focus of the next two articles.

Here’s a quick summary of the articles I am planning in the series. My goal throughout will…

INTUITIVE AUDIO DEEP LEARNING SERIES

A Gentle Guide to the world of disruptive deep learning audio applications and architectures. And why we all need to know about Spectrograms, in Plain English.

Image for post
Image for post
Photo by Jason Rosewell on Unsplash

Although Computer Vision and NLP applications get most of the buzz, there are many groundbreaking use cases for deep learning with audio data that are transforming our daily lives. Over the next few articles, I aim to explore the fascinating world of audio deep learning.

Here’s a quick summary of the articles I am planning in the series. My goal throughout will be to understand not just how something works but why it works that way.

  1. State-of-the-Art Techniques— this article (What is sound and how it is digitized. What problems is audio deep learning solving in our daily lives. …

INTUITIVE TRANSFORMERS SERIES NLP

A Gentle Guide to the inner workings of Self-Attention, Encoder-Decoder Attention, Attention Score and Masking, in Plain English.

Image for post
Image for post
Photo by Scott Tobin on Unsplash

This is the third article in my series on Transformers. We are covering its functionality in a top-down manner. In the previous articles, we learned what a Transformer is, its architecture, and how it works.

In this article, we will go a step further and dive deeper into Multi-head Attention, which is the brains of the Transformer.

Here’s a quick summary of the previous and following articles in the series. My goal throughout will be to understand not just how something works but why it works that way.

  1. Overview of functionality (How Transformers are used, and why they are better…

INTUITIVE REINFORCEMENT LEARNING SERIES

A Gentle Guide to the REINFORCE algorithm, in Plain English

Image for post
Image for post
Photo by Michael Dziedzic on Unsplash

This is the sixth article in my series on Reinforcement Learning (RL). We now have a good understanding of the concepts that form the building blocks of an RL problem, and the techniques used to solve them. We have also taken a detailed look at two Value-based algorithms — Q-Learning algorithm and Deep Q Networks (DQN), which was our first step into Deep Reinforcement Learning.

In this article, we will continue our Deep Reinforcement Learning journey and learn about our first Policy-based algorithm using the technique of Policy Gradients. …

INTUITIVE TRANSFORMERS SERIES NLP

A Gentle Guide to the Transformer under the hood, and its end-to-end operation.

Image for post
Image for post
Photo by Joshua Sortino on Unsplash

This is the second article in my series on Transformers. In the first article, we learned about the functionality of Transformers, how they are used, their high-level architecture, and their advantages.

In this article, we can now look under the hood and study exactly how they work in detail. We’ll see how data flows through the system with their actual matrix representations and shapes and understand the computations performed at each stage.

Here’s a quick summary of the previous and following articles in the series. …

INTUITIVE REINFORCEMENT LEARNING SERIES

A Gentle Guide to DQNs with Experience Replay, in Plain English

Image for post
Image for post
Photo by Josh Riemer on Unsplash

This is the fifth article in my series on Reinforcement Learning (RL). We now have a good understanding of the concepts that form the building blocks of an RL problem, and the techniques used to solve them. We have also taken a detailed look at the Q-Learning algorithm which forms the foundation of Deep Q Networks (DQN) which is the focus of this article.

With DQNs, we are finally able to being our journey into Deep Reinforcement Learning which is perhaps the most innovative area of Reinforcement Learning today. We’ll go through this algorithm step-by-step including some of the game-changing…

INTUITIVE TRANSFORMERS SERIES NLP

A Gentle Guide to Transformers, how they are used for NLP, and why they are better than RNNs, in Plain English. How Attention helps improve performance.

Image for post
Image for post
Photo by Arseny Togulev on Unsplash

We’ve been hearing a lot about Transformers and with good reason. They have taken the world of NLP by storm in the last few years. The Transformer is an architecture that uses Attention to significantly improve the performance of deep learning NLP translation models. It was first introduced in the paper Attention is all you need and was quickly established as the leading architecture for most text data applications.

Since then, numerous projects including Google’s BERT and OpenAI’s GPT series have built on this foundation and published performance results that handily beat existing state-of-the-art benchmarks.

Over a series of articles…

INTUITIVE REINFORCEMENT LEARNING SERIES

A Visual Guide to how and why the Q Learning Algorithm works, in Plain English

Image for post
Image for post
Photo by Carlos Esteves on Unsplash

This is the fourth article in my series on Reinforcement Learning (RL). We now have a good understanding of the concepts that form the building blocks of an RL problem, and the techniques used to solve them. We can now bring these together to learn about complete solutions used by the most popular RL algorithms.

In this article, it is exciting to now dive into our first RL algorithm and go over the details of Q Learning! …

INTUITIVE REINFORCEMENT LEARNING SERIES

A Visual Guide to techniques used by Value-based and Policy-based solutions, in Plain English.

Image for post
Image for post
Photo by Aidan Howe on Unsplash

This is the third article in my series on Reinforcement Learning (RL). Now that we understand what an RL Problem is, and the types of solutions available, we’ll now learn about the core techniques used by all solutions. Using an iterative algorithm as a framework to incrementally improve predictions, we’ll understand the fundamental similarities and differences between Value-based and Policy-based solutions.

Here’s a quick summary of the previous and following articles in the series. My goal throughout will be to understand not just how something works but why it works that way.

  1. Intro to Basic Concepts and Terminology (What is…

Ketan Doshi

Machine Learning and Big Data

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store