Sign in

Lead Data Scientist Consultant || Ex Co-Founder & CTO of a funded AI startup

Question generation using state-of-the-art Natural Language Processing techniques

Image by Author

Question answering is a very popular task in Natural language processing but question generation is novel and hasn’t been explored much yet.

If you want to try a live demo of question generation in action, please visit

Question generation has a lot of use cases with the most prominent…

The most practical use of word embeddings (word2vec, glove, etc) you will ever see.

Image by Author

King — Man + Woman = Queen

You might have seen the traditional word2vec or Glove word embeddings examples that show King -Man+Woman = Queen. Here Queen will be returned from the word embedding algorithm given the words King, Man, and Woman. Today we will see how we can use this structure to solve a real-world problem.

1. The problem definition:

Create a docker container for summarization task on a GPU

Image by Author using a free stock image from

Using GPU within a docker container isn’t straightforward. There shouldn't be any mismatch between CUDA and CuDNN drivers on both the container and host machine to enable seamless communication.

The ideal approach is to use NVIDIA container toolkit image in your docker that provides support to automatically recognize GPU drivers…

Open-sourcing paraphraser trained on a custom dataset and T5 large model

Image by Author


The input to our program will be any English sentence -

Four private astronauts launched to orbit by Elon Musk’s SpaceX returned to Earth Saturday evening, splashing down into the ocean off the east coast of Florida after a three-day mission.


The output will be paraphrased version of the same…

Group semantically similar documents easily

Image by Author

In semantic search, we search a database of documents for a given user query and get back a set of relevant documents.

Whether we are building a recommendation system or question answering system often times we get back duplicate search results that are not exactly identical to be called duplicates…

Tips to market your AI course

Image by Author

This year I launched my first ever online course on Udemy titled “Question generation using Natural Language processing” and made $3333 dollars in 5 months with 0 marketing spend and 100% paid enrollments.

A practical AI project using HuggingFace transformers and Gradio App

Image by Author

Practical use case

Using Wordnet, Conceptnet, and Sense2vec algorithms to generate distractors

Image by Author

What are distractors?

Distractors are the wrong answers in a multiple-choice question.

For example, if a given multiple choice question has the game Cricket as the correct answer then we need to generate wrong choices (distractors) like Football, Golf, Ultimate Frisbee, etc.

Minimalistic code for few-shot text generation with HuggingFace

Image by Author

I’m sure most of you have heard about OpenAI’s GPT-3 and its insane text generation capabilities learning from only a few examples.

The concept of feeding a model with very little training data and making it learn to do a novel task is called Few-shot learning.

A website GPT-3 examples…

A Google Colab code notebook exploring a real-world use case with OpenAI CLIP

Image by Author

Problem Statement:

Imagine that you are a writer and you are searching for the best image that goes with your blog or book. You have a search phrase in mind like “Tiger playing in the snow”.

Ramsri Goutham

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