Project Overview
- This project I would say, was the birth of the curious, analytics driven researcher in me
- In India, most folks make a choice between engineering and medicine after high school. While my choice was with engineering, a part of my mind was with medicine as well at times
- This project allowed me to envision the wide reaching applications and implications engineering and tech can have across domains, including the medical field
- The idea behind this project was to replicate the experiment for a guilty knowledge test polygraph, but instead do it with brain waves
- This involved understanding a fair amount of what brain waves (frequencies) show up during various brain functions, draft up guilty and non-guilty version of stories exposed to our recruited candidates
- Also since we were not really funded, the only EEG apparatus we had access to was one in a government hospital requiring pasting the electrodes to the candidate's scalps
- While we were not able to prove our hypothesis on clear distinctions between guilty and innocent knowledge of a scene / story from brain waves, with satisfactory model metrics, we released the dataset here with the positive class denoting guilt
Skills
Experiment setup
A part of the low success chance / noise of this project, was due to the constant distraction of patients and staff around the experiment setup in the public hospital. This was a chance to learn that execution is as important as planning
Applied analytics
For something we started out of a wild idea, it was highly satisfactory to see the plan come to life, giving me the inspiration to pursue applied analytics to a wide variety of domains such as finance, compliance, healthcare and more...
The world of publishing
Before I would discover the world of shipping ML into product, this was my gate to explore the world of pushing boundaries in ML and getting it out into the world through publishing papers, datasets, ontologies and even patenting