2023 Year in Review

28-December-2023

As the year 2023 is coming to an end, it is time to reflect on what the previous twelve months have been like and reflect on my goals for this year. This year has made me realize that some chapters of our journey don't unfold as planned (and that maybe I am lazier than I prefer to think of myself). The echoes of unfinished aspirations serve as sad reminders of the unpredictable nature of our paths and the humbling truth that not every goal is supposed to be achieved in our predetermined destiny.

First Quarter

14th of January

I navigated around my computer, collected all of the old Python projects, and uploaded them all into one single private GitHub repository.

6th of January - 26th of March

These were the only 80 days of the Year I did Journaling before bed, adding a little more, I only did journaling in 34 of the days in these 80 days.

23rd of March - 29th of March

These days, I replicated my portfolio website's desktop design into a Figma design and tried to look for some spots where I could get it a bit improved. The changes that I made in the design were in the Blog Section, the Project Section, and also a few more things, you can read more about it.

25th of March - 31st of March

I tried to implement the things I learned from working on a Physics Engine into a new project. So I made a Basic Gravity Simulator and just added a Basic Collision Response system for Circle to Circle. I also added a way so that I can add more Objects to the simulation while it is running. In the end, I added a Star (basically an Object with a big mass) and added a bunch f small objects orbiting around the star in 2 different rings.

Second Quarter

15th of April - 1st of May

I tried working on the Separating Axis Theorem or SAT for the Physics Engine. SAT is an algorithm in which we measure the angle of every edge of both the Closed Convex Polygons in being detected, and Mathematically assume an axis in that angle and in the perpendicular and check if both the Polygons are separated in the axis or not. If they are separated in any one of the axes, we say these Polygons are not colliding.

Third Quarter

8th of September - 27th of September

I also built my first-ever Neural Network, which was trained and tested on the standard MNIST Dataset (Modified National Institute of Standards and Technology dataset), a dataset of thousands of handwritten digits on 28 by 28 pixels. The Neural Network had 784 Input Neurons, 10 Neurons in the Hidden Layer, and 10 Neurons in the Output Layer.

I also have a blog about Working on my First Neural Network.

Fourth Quarter

1st of October - 17th of October

I also tried to expand the Neural Network from 2 Layers to 3 Layers, 2 Hidden Layers with 16 Neurons each and 10 Neurons in Output Layer. I made changes to all the functions working on Layers (Initializing Parameters, Forward-Propagation, Back-Propagation, and Updating Parameters). As I tried to Train the Network and was being shown its accuracy, it was fluctuating around instead of gradually climbing. I knew there was some problem with the Back-Propagation Function or it was supposed to be written a different way maybe, but I still haven't figured it out.

I also wrote about it.

20th of December - Still

I am working on a small project, which I uploaded into a new Private GitHub Repository on the 23rd of December. I don't want to share a lot about it right now, but it has something to do with Mathematics.

Conclusion

Reflecting on the Goals

Goals for 2024