There are lots of different benefits to trying to replicate computational biology scientific papers.
First, it's fun. Yeah, that would be enough. I, an inexperienced aspirant researcher, trying to understand, explain, and replicate the work of some great scientists. At least it's gonna be an adventure - and maybe a little frustrating considering my limited computing resources.
Second, I need to learn. I won't do anything relevant in the field without exposing myself to some real work, from who is actually in the arena.
Third, I do not come from a fancy research university like an Ivy League. Don't get me wrong, Insper is great. But if I want to compete with the folks from these schools to be accepted into a Ph.D. program, I need to do something (maybe lots of things) to stand out.
I'm still not sure if this blog is going to be an adequate channel for publishing such content. But if turns out it's not, I will just copy every post and post on https://medium.com/.
How this is going to work:
First, it's fun. Yeah, that would be enough. I, an inexperienced aspirant researcher, trying to understand, explain, and replicate the work of some great scientists. At least it's gonna be an adventure - and maybe a little frustrating considering my limited computing resources.
Second, I need to learn. I won't do anything relevant in the field without exposing myself to some real work, from who is actually in the arena.
Third, I do not come from a fancy research university like an Ivy League. Don't get me wrong, Insper is great. But if I want to compete with the folks from these schools to be accepted into a Ph.D. program, I need to do something (maybe lots of things) to stand out.
I'm still not sure if this blog is going to be an adequate channel for publishing such content. But if turns out it's not, I will just copy every post and post on https://medium.com/.
How this is going to work:
- Every month I'm going to pick a computational biology paper and replicate it;
- Then I'm going to write a blog post explaining what I've done, my biggest challenges, and my learnings;
- Finally, I'm sharing it with you, my mysterious reader.
The first paper's title is "Diffusion probabilistic models beat GAN on medical 2D images". You can find the paper here. If you are not from the machine learning field, you may get scared by this fancy name. It is just saying that two methods are going to be compared on how great they can generate synthetic medical images.
I picked this one because I'm currently on the computer vision course at Insper. Practicing my skills with images (some type of medical data I have barely worked before) is going to compose with what I'm learning in class.
Pray with me for Insper to allow me to use its high-performance computer, because mine won't be enough.
Soon (I hope) I will be back with the result.
Wish me luck, bye!