Chapter 0: Introduction to epsilon

Hi. I’m Epsilon, and this is a blog.

I’m a Computer Engineer, specialised in Machine Learning and Computer Vision. Most of my research works centers around automated detection in Archaeology. I have introduced the use of YOLO in LiDAR detection for this purpose, in a research internship at the University of Leiden. I have also worked in rare event detection in train borne systems at the SNCF-CIM. I also have quite the experience in social media intelligence, i.e. real-time (or close to) analysis of very large amount of social media data (NLP and CV tasks).

I am very passionate about Mathematics, especially Discrete Dynamical Systems and Chaos. I find that those fields lend themselves very easily to Computer exploration and programming. Although the maths behind fractals can be quite spicy, the computer code is usually not, and so my point of entry in those domains was found. While many of my projects centers around Mathematics, they are not limited by it. I also try out a lot of things with Deep Learning, in particular Generative Adversarial Networks, to try and create interesting looking things. Finally, I love Music, the hard hitting kind and the more soft weird classical kind. I like to mix those three things.

My main programming languages are Python, C and Processing. I have used Processing quite extensively to make my animations, although I tend to port things in C when it requires too much compute power. My Deep Learning projects are usually in Python, simply because of the enormous amounts of frameworks available. While I do like to reinvent the wheel sometimes, having someone else write the code for backpropagation does feel good :)

While I often post stills and animations on specialised groups on Facebook or Instagram, I don’t usually explain exactly what’s going on; so this site acts as a place where I can write in more details about what I’m working on.

All in all: I try to make things that looks like art, but are just greyscale Mathematics if you look at it closely.

See ya.