Carnatic Music features rich recurrent patterns from a few building blocks — making it a rich playground to explore AI based music generation. Let’s see how LSTM can generate new basic compositions — Building Blocks in Carnatic Music Carnatic music is one of 2 Indian classical music forms (the other being Hindustani music). Unlike in Western classical music where compositions are set to a particular key (e.g. Beethoven’s Symphony №5 in C minor) and features multiple modes; Carnatic music compositions are mostly set to a single distinct mode…

A spectrogram analysis of the Carnatic version of the C Major scale — The basics Sankarabharanam is the Carnatic version of the major scale (or Ionian mode). Here are the notes of Sankarabharanam: Ascending: S R₂ G₃ M₁ P D₂ N₃ Ṡ Descending: Ṡ N₃ D₂ P M₁ G₃ R₂ S This is the 29th Melakarta raga in Carnatic music. As I explained in this…

A deep mathematical and musical exploration into the Indian Carnatic Music Raga Bhairavi that is older than 1,500 years. I give key examples of patterns and combinations used during Carnatic jazz style musical explorations of Bhairavi. — There are a total of 72 whole scales (also called Melakarta Ragas) in Indian Carnatic classical music, as I’ve gone through in a previous blog. In this article I’ll talk about one raga in Carnatic music called Bhairavi (different from the Hindustani Bhairavi) which is not one out of the…

One day I realized it was time for a new adventure. Here’s how small consistent efforts laid the foundations for my next role. — I want to share my experiences in transitioning from academia to data science as it might benefit many in a similar boat. More broadly, my experiences could also help lay a solid foundation for achieving your goals —in transitioning to a completely different field. I’ve always wanted to be a…

The classic deep learning for computer vision example project starts out with a dataset containing images, and labels. However, when computer vision is needed for solving business problems, data is usually unlabeled and labeling the data is itself a challenge. This article walks through labeling images at scale and associated challenges — The classic example of a deep learning computer vision project starts out with a dataset containing images, and labels. Depending on what type of labels you have and the task you want to accomplish — image classification, object detection, or image segmentation; you would select from an appropriate set of…

OSMnx is a powerful Python package for road networks but does not incorporate current traffic conditions — let’s use the HERE traffic API to impute real-time traffic information in OSMnx — OSMnx is a powerful Python package for static road network information taken from OpenStreetMap- a free editable geographic database of the world. Here is a brief code snippet showing how OSMnx can be used to obtain the road network of Washington DC:

Tutorial on transfer learning using the Resnet-18 deep learning CNN to achieve a 97% validation accuracy — Deep Convolutional Neural Networks applied to image prediction tasks are extremely popular. It seems as though every few months there is a new algorithm that pushes the limits on CNN architectures. Just a few years ago these state of the art algorithms and the knowledge behind how to build them…

With Amazon announcing drone super-highways, it becomes all the more important to consider cyber-physical security in drones — What comes to mind when someone mentions drones? For me it’s one of two things: toy drones (like the one we recently gave our friend’s kid who turned 6) — or military drones and unfortunately, images like this:

SafeGraph provides dynamic location based data. Let’s use this to derive research insights. — On their website, SafeGraph mentions they have over 11M+ records for Points of Interest (POI) including data on location, brands, etc. Looking at the documentation, there are basically 3 kinds of data sets: places, geometry, and patterns (they recently introduced a new Spend data set as well). …