Musings

Here, I write out random thoughts and reflections on arbitrary topics. I try to keep my thought sequence as coordinated as I can. But that’s a feat rarely accomplished by even the most articulate writers. And in as much as I aspire to acquire the ability to articulate my thoughts into concise, crisp sentences, I am without doubt no where close to ideal. And I’m not being modest. So strap tight your seat-belt and brace yourself for an unceremonious immersion into the murky depths of my befuddled thoughts. Godspeed comrade!

 

  • On Convex Optimization
    • In this post, I talk about Convex Optimization and the various algorithms for solving Unconstrained and Constrained Convex Optimization problems.

 

  • On Optimization
    • In this post, I talk about Optimization as a problem-solving approach and the various flavors of optimization problems.

 

 

 

 

  • On insect navigation
    • In this post, I write about the various strategies insects use to navigate to and from food sources as well as general lessons we could learn from the insect species with regards to specialization and resource efficiency.

 

 

 

  • On Semantic Robot Programming
    • In this piece, I talk about the significance and modus operandi of Semantic Robot programming as well as some questions I have about it.

 

  • On Skill Discovery through Skill Chaining
    • In this post, I briefly talk about a skill discovery algorithm by Professor George Konidaris and Professor Andrew Barto and about some questions I have about this work.

 

 

 

  • A primer on KL Divergence
    • This post is a brief primer on KL Divergence. It is an answer to the first of my list of confounding questions.

 

 

  • On Bayesian Optimization
    • In this post, I think about the motivation, workings and implementation of Bayesian Optimization.

 

 

  • On Learning to Guide Task and Motion Planning
    • In this post, I deliberate on current efforts to apply learning in accelerating task and motion planning for robots. Works discussed here are mainly by MIT PhD candidate Beomjoon Kim from the Learning and Intelligent Systems Lab at MIT.

 

 

  • Reinforcement Learning Notes
    • Here, I write quick notes about RL that I find necessary to aid my understanding or to guide my thoughts about certain concepts in RL.