Artificial Intelligence workshop

Humankind never stops innovating, and one of the technologies that is driving innovation is artificial intelligence. Artificial intelligence is predicted to cause a paradigm shift in the way medicine and healthcare works . Since it’s never a bad idea to get yourself upto date with the latest technologies, I decided to participate in a 10 day AI workshop named Inspirit AI scholars conducted by Stanford University alumni.

The whole program was really interesting, and I learnt many new concepts related to AI. One of the highlights of the program was the environment in which it was conducted. I got to interact with students from different countries and was instructed by Stanford alumni who were not only very learned in their spheres of knowledge but also treated us like fellow colleagues, answering every question we had related to AI and even college life. The program wasn’t all studying – the instructors had some really exciting games that helped us to bond and get working as a team.

The first few sessions were concentrated on learning the basics of python, the programming language that we were going to use. During that period we learnt about different concepts that we would integrate in our projects that we had to present towards the end of the program.

For me, the most interesting part of the program was the project – my team and I made a project on how to detect Pneumonia from chest X-ray scans. We used a number of models in python to predict whether a lung had Pneumonia or not. We used a type of Machine Learning known as Supervised Learning to do this (Supervised Learning means that the machine finds certain patterns and makes predictions based on a set of input and output data). Here’s how we went about designing and executing our project-

For starters, we had a set of 2400 chest X-rays that we had to sort into testing (400 images) and training data (2000 images). Our input was the 2000 images, while our output was the predictions made by the program. The training data images had labels on them – we essentially told the program that this image had Pneumonia while this image didn’t have Pneumonia. The program created its own rules regarding the detection of Pneumonia (chest X-rays with Pneumonia appear darker due to a build up of fluid in the lungs – this could be one of the parameters that a program could use to differentiate between not Pneumonia and Pneumonia). The rest 400 testing data images were the images that we would feed into the program to see how well our models worked (basically, how correct their predictions were).

Some of the models that we used were logistic regression (given by the function Log), K Nearest Neighbours (knn) and decision tree classifier (Dt). These three were comparatively primitive models that gave us accuracies of 68-72 percent. More complex models that we used were Convolutional Neural Networks and Multi Layered Perceptrons , the MLPs giving us 83.35 percent accuracy.

This made me realize that by using better models and using more amount of data the predictive power of the program increases. AI is a very powerful tool in various fields specifically medicine, where certain tasks that are currently performed by doctors can be performed by AI in the future. Perhaps deficiency of doctors can also be countered by integrating AI into healthcare, providing faster and accurate diagnoses.

Overall, the whole workshop was an amazing experience. During the program we started with the basics of programming, introduction to AI and within a short period of time working with our project teams and instructors made our own AI application based projects. My project in the area of healthcare application of AI opened doors of curiosity for me. Apart from the technical aspects, I had an opportunity to hone my soft skills, like communication, working in cross cultural teams and problem solving.

I would like to thank Mr. Adeesh Goyal and Mr. Artem Trotsyuk (my instructor) of Inspirit AI for organizing this workshop along with all the instructors who taught me so much and made this experience one that I shall remember for a long time.