(Attempted) All Nighter Productivity Log

(Attempted) All Nighter Productivity Log

March 22, 2024
Productivity Log, productivity

(Attempted) All Nighter Productivity Log #

I am going to stay up as long as I’m able to be productive. I am working in the ISC with Michael sitting a few feet behind me, so he’s able to see my screen. I hope that this will stop me when I am tempted to stop working.

Tasks #

  • Implement graphing visualization for recursive feature elimination of side channel leakage dataset
  • Set up programmatic email sending with Gmail and custom domain for Griffin Events
  • Do Project 3 for Applied Machine Learning, which involves implementing a selection of machine learning classifiers
  • Upload my updated website once the ftp server is back up

Log #

21:16 - I begin working on my research project.

21:33 - I added scikit-learn as a submodule of our repo so I can modify it to extract the data I need that is internal to the RFE process for the visualization.

21:34 - I have to compile scikit-learn and it’s taking a while.

21:41 - sftp works again on my web server.

21:43 - scikit-learn is an impressive piece of software. It’s very well written.

21:56 - I’m fixing a couple links on my site, then it’ll be ready to ship.

22:18 - We’re evaluating what db we want to use for Griffin Events. We’ll probably go with firestore.

22:19 - Uploading the site now.

22:23 - Uploaded the site. Nice to finally have it up. Back to focusing on research.

22:40 - I added the graphing functionality to RFE in scikit-learn.

23:50 - things happened, a girl named after a greek goddess left her car keys in my backpack, and now we’re working again in Sadler.

23:51 - Just downed a full cup of coffee.

00:05 - I found an email service that lets you send 300 emails a month for free. This’ll be what we use at first. I set up the DNS records on our domain.

00:07 - I love life, anything is possible with software, I love being excited and being optimistic, have wagmi energy or gtfo!

00:24 - Sent test email. We will use this service until we deploy griffin events, then we’ll probably switch to gmail or sendgrid.

00:28 - apparently sadler is closing?? we gotta move somewhere.

00:36 - Michael’s worn out, so he went home. I’m wired as all hell, so I have an obligation to the universal energy to continue working.

00:37 - I’m going to listen to the soundtrack for The Social Network. It’s an auditory simulant. Sometimes weird things like music just help me unlock my brain and accomplish everything all at once.

01:21 - I finished implementing the email api for griffin events! It works really well. I had to learn what the RFC 3339 format was.

01:23 - I need to fix a bug in the RFE visualization. The graph isn’t clearing on each iteration and the results from each classifier are being overlaid on top of each other.

01:33 - Fixed! The results are pretty goofy. By removing a few features, the accuracy actually goes up! and there are some straight up anomalous results for some of the classifiers where they’ll just decrease in accuracy then go back up as features are removed.

01:34 - On second thought, those changes in accuracy are really small now that I look at the scale of the axes. That’s within an expected tolerance of error. Nice, I’m done with that now!

01:38 - Okay, I pushed the code. Apparently when you make changes to a git submodule, you have to enter the top level directory of the submodule, commit the code, then commit and push again in the main repo.

01:39 - I’m going to work on my Applied Machine Learning project now. From what I remember from skimming it two weeks ago, it’s just implementing a bunch of classical ml algs with scikit-learn. easy peasy.

01:52 - Can someone remind me how calling functions with parameters given to me by the professor with zero comprehension is going to teach me machine learning? Oh, it won’t teach me anything? Ohh, thanks for telling me.

1:57 - Alright, I’m starting to struggle to focus. If I’m not being productive, I have no right to be awake. gn :huggingface:

Accomplished #

  • created dataviz for research, learned more about scikit-learn internals and git submodules
  • setup email service, learned more about email campaigns + A/B testing
  • first half of p3 for aml, need to validate answers
  • uploaded website, fixed links. it’s all pretty now! i still need to figure out why the images on the gallery aren’t loading before I publish it.