The safety net, in a competitive society is healthcare.

From my observations as a kid I have seen in my family and relatives struggle to find a matching doctor, a clinic or a hospital when growing up in India. Failures in efficiently navigating the healthcare market is common. It is demanding on money, social support and time from people who cannot afford it.

In this healthcare setting, mental health is a scorned topic.

This discussion is based on data from data.gov.in, however the issues are true across geographies. …

I frequently ran into conflicting opinions and thoughts, when discussing or understanding people I don’t know. These people I don’t have personal relationships with. With those I do, I have shared experiences, and I slowly learn how to live with them.

A common symptom that I observe when I rationalize decisions or actions of the persons I don’t know, is my ignorance of detail. This ignorance leads to generalization. I first generalize an expectation, I have about those persons. An expectation from something unrelated I know about them, like country, nation of origin, city, state, list goes on. Because I…

What is a photograph ?

Imagine a scene. Freeze a moment when the scene is at the height of its narrative arc, capture it with a camera. The decisive point. This moment selected is put on a medium. The medium transforms the moment snapped into a visual experience. Thus a unique point in someone’s timeline where something happened is captured in one of many ways inside a photograph.

How does one pick a fictional scene ?

Photographer McFarland is a maker. Combination printing is a technique of joining multiple negatives into one continuous photographic fiction. McFarland practices the technique with large format films. …

I have notes here for the exercises I have completed from Sutton’s book on Reinforcement Learning. The techniques picked from here are applied to environments in the OpenAI gym which are documented on this link.

Chapter 2

This chapter discusses a test environment. An N-Armed Bandit test-bed for RL agents.

The test-bed takes as input the `The Agent` definition. The test-bed runs `G` repetitions of `P` plays of the agent against the N-armed bandit. In the book the term `task` is used instead of `game`, which I have used in my code and notes.

You can find detailed description of N-armed bandit…

Today I snatched some time out of my day to peruse gym.OpenAI.com/docs. Earlier in the year I had taken a stab at reinforcement learning by buying Richard Sutton’s book from Amazon. The first two chapters had me hooked on and I solved exercises for the first one. Moving on from second to third was difficult because I skipped exercises from chapter two then I had no problem to ponder on when I was not reading. It was not as fun solving problem sets in my free time. This lead to a full stop in my exploration of RL. I had…

Finding Amino

For a few weeks I had to put myself through conscious questioning to rationalize my next transition and why I want it. The task is daunting and the outcome was sprinkled with compromises.

I like to believe somehow my ambitions give a long term direction to my curiosities and that sets the premise for such decisions.

I am always curious and excited to work on problems in technology companies in health, biosciences and education. I was researching about startups in this field. The internet was useful. I came across this post by David Vivero while randomly sky-jumping…