Anjishnu
Anjishnu
CS PhD student. Working on Fairness and Interpretability methods in NLP.

Week Three

Week Three
Wanderlust 🥺 | Teamlab Borderless Museum by Luke Paris on Unsplash

I had multiple exams this week at my university along with some hard final deadlines for assignments and stuff. So, I couldn’t actually post this blog on time, however, I am still writing it later on as a week3 blog, so as not to disturb the continuity.

This week, my major contributions included working towards correcting the remaining loss functions in mlpack. I finished writing demo notebooks to compare the new implementations with the PyTorch implementation. All the notebooks are available in a demo repository that I am maintaining here. I believe this notebook based approach towards features that share similar behaviour with other popular libraries would help not only towards a more robust review process and also towards beginner adoption of the library for common tasks involving the ANN module. There was an official mlpack meeting this week, which I missed because I had an exam early the next day. However, I had informed my mentor previously about the issues with the loss functions. And he raised the point during the meeting on my behalf about having a stricter review process in the future to have fewer possible bugs, with my Google Colab notebook based approach that I am taking, being suggested as a possible alternative.

One cool thing that happened during this week. I noticed that a lot of people cloned my demo repository for the new mlpack features that I am working on during the last two weeks. That was a major feel-good factor for me.

Not much happened other than that for this week. So, coming to the end of this blog, here is the song for the week.

The Kid I Used to Know by Arrested Youth

See you next week!
XOXO