My Blog

So I made a blog. I am planning to write mainly about my projects but also about interesting things from Machine Learning, Robotics and other areas I come across.

Wayve - spatial pooling attention for feature selection

When working at Wayve, for the second half of the internship I had my own project, adding an attention based pooling layer to add interpretability into our self-driving model and improve results by more densely taking information from more relevant areas of the image.

Hazard label detector with YOLOv3 and automatic dataset generation

A part of RoboCup Rescue competition focuses on sensing, one of the tasks is detecting various objects in video, a lot of them rather obscure. We had to create our own dataset, which I did by generating it by placing the planar objects transformed onto background images. Then I trained a YOLOv3 on the generated data and tested it on real life video.

Book review and recommendations

A selection of the most intersting books I've read with recommendations based on which area you'd like to learn more about.

Leadership, book review, hackathons, student societies and other stuff

A summary of leadership books I've read, lessons I've learnt and how I have applied them in hackathons and student societies.

Scaling hyperparameters with batchsize

An exploration of a random idea that if you scale various hyperparameters correctly, you should be able to scale batch size without affecting results.

NEAT Mars Lander

This summer we got a task of programming a mars lander simulation and autopilot. Using a ready made mars lander gui, we were to add simulation and autopilot. I decided to extend the exercise by training an autopilot using Neuro-evolution of augmenting topologies (NEAT) and training it to efficiently land in many scenarios.