Clear, simple and useful NLP blogs.
By someone who loves NLP, writing and teaching.
How does automatic differentiation really work?
Frameworks like PyTorch and Tensorflow allow us to perform easy deep learning experiments because we don’t have to worry about backpropagation. However, ever wondered how do these frameworks actually perform differentiation? This article will help you understand the internal working of automatic differentiation.
Customize transformer models to your domain
Transformers like BERT are pre-trained on a wide range of corpora, but can they achieve similar performance feats on specific domain related tasks? Learn how to customize transformers like BERT, RoBERTa to adapt to specific domains like healthcare, finance.
How to ace your take-home data science assignment?
Follow these steps and ensure that your data science job assignment doesn't look amateur.
BERT-ology at 100 kmph
Among different BERT models developed by Machine Learning researchers, many of them have used interesting ideas that tackle problems common to a lot of Machine Learning, NLP tasks, and not just transformers. It is important to understand how and why did they work!
How to avoid getting a data science job
It is hard to get a data science job. It is even harder to escape a scam job. In this post, I share practical advice for aspirants starting out their data science career. Especially, how to be selective about a data science role at startups.
Graph Convolutional Networks for dummies
Graph Neural Networks powerful tools used by large companies to make recommendations, link financial transactions, detect fraud behavior. This post explains how deep learning has enabled a powerful understanding of graphs.
Explain a prediction using LIME
LIME is a framework that can explain any Machine Learning model by training a secondary model around the point whose prediction is to be explained.
If you wanted to appreciate someone today, now is the time.
Sign up for my newsletter and you’ll never miss a post.