And it’s that time of year again! Most of you must have started planning for a brand new year and a new decade. Keeping up with that spirit, this post is about my plans and expectations regarding this blog space for the upcoming new year.
Its been almost 5 years since I started this blog. Those of you who have been following me here, would be aware, this has not been easy to follow at all. With frequent changes in the design, theme, web address (from github pages to datasciencevision.com to the current reckoning.dev!) and underlying technological changes (pelican to nikola to hugo to gatsby!). The very first promise I want to make in the 2020 is to guarantee the stability of this space.
- reckoning.dev address is going to stay for long, its been paid for the next 5 years!
- The current looks/theme (thanks to the work of Tania Rascia) along with the underlying tech of gatsby, mdx and react is something I have fallen in love with.
2019: A Brief SummarySection titled 2019: A Brief Summary
A year is a long time. Still, when I look at my blogging history, this has been the year with one of the lowest number of posts. Unfortunately - or fortunately - I’m learning so much, so fast, that I haven’t been able to write about it all. But at least I can track it here, and try my best to get a resource out into the world for my own reference and to help others. Following is a brief summary of my 2019:
Top 5 things that I learnt this year:
- javascript, react, gatsby, css etc.
- mlflow for managing deep learning experiments
- mixed precision training in pytorch
- Writing user-defined algorithms as pytorch modules in CUDA
- Designing complete python library: Developed a pytorch based high level training/inference engine to train models with medical imaging data.
Top 5 things I developed this year:
- Pytorch based high level library for training deep learning models on medical imaging data.
- Flask apps to create APIs for controlling SLURM jobs via web.
- Model zoo API and database using mongodb for management of models.
- Command line tools to submit and manage jobs on a compute cluster.
- Weakly supervised learning models for breast cancer.
Top 5 Books I read:
- Bad Blood: Secrets and Lies in a Silicon Valley Startup by John Carreyrou
- The Emperor of All Maladies by Siddhartha Mukherjee
- Brief Answers to the Big Questions by Stephen Hawking
- The Patient will See you Now by Eric Topol
- The Master Algorithm by Pedro Domingos
My Incomplete Tasks and Failures:
- Object detection for extremely imbalanced data and loose bounding boxes
- Generative models for breast mammogram data
- A GUI for debugging of deep learning models on medical images
- A markdown editor/preview app for OSX
2020: New year, New PromisesSection titled 2020: New year, New Promises
With the beginning of a new year, I want to make few new promises to myself. In no particular order, following are list of things I want to do/learn/complete in this year.
Things to do for this Blog:
- Write Reviews of Books I have read:
- The Patient will See you Now by Eric Topol
- Bad Blood by John Carreyrou
- The Emperor of All Meladies by Siddhartha Mukherjee
- The Gene by Siddhartha Mukherjee
- Brief Answer to Big Questions by Stephen Hawking
- Deep Medicine by Eric Topol
- Language Intelligence by Joseph J. Romm
- Problem Solving 101 by Ken Watanabe
- The Master Algorithm by Pedro Domingos
- Life 3.0 by Max Tegmark
- How to lie with Statistics by Darrell Huff
- Sapiens: A Brief History of Humankind by Yuval Noah Harari
- Why We Sleep by Matthew Walker
- Einstein by Walter Isaacson
- The Signal and the Noise by Nate Silver
- Origin Story by David Christian
- The Grid by Gretchen Bakke
- Abundance by Peter H. Diamadis and Steven Kotler
- Write about major Research Papers I read
- Write in-depth about some of the gatsby topics
- Write more about some of new deep learning algorithms and their implementations
- Write more about travel!
Things to Learn and Develop:
- VSCode Extension Development
- Contributing to Pytorch Code base
- Get involved with the Detectron 2 codebase for Object detection
- Advanced CUDA Programming
- React + Django App Development: A Markdown Editor
- Getting Familiar with Theme-UI and convert this Blog’s CSS to Theme UI
- Release theme of this Blog as a stand-alone gatsby theme
- Bayesian Deep Learning
- Reinforcement Learning
- Publish at least 3 papers
Few Health Goals
- Do Yoga at least 5 hours a Week
- More Hiking
- Visit at least 20 new Beaches
All in all, 2019 was a great year, and I learned a lot! It went by too fast. I’m excited for 2020. Thank you all for reading, contributing, discussing and following my content here. I hope to be a lot more productive here in 2020. Please do follow me on Twitter
COMMENTS