Some of my stuff and experimentations I have been doing in the fieds of AI, scripting and automation
The internet is great, we can actually find information about any subject. The difficult part is then to organize them and classify those data in the right bucket given the topic you are interested in. A classical approach would be to go through everything, read it, store the topic you extracted somewhere somehow (in your head, post-it, csv file…), and then classify manually all those documents in the right bucket. It can be somehow ok if your corpus is not too big, but once you reach a dozen documents whose length may vary a lot, it becomes quite a burden. Moreover, you may be interested in only one of the topic covered in the document and would still have to read the whole bunch and your classification may not be consistent. Surely there is a better way to use our time…
Machine learning is great but it is also a difficult and somewhat cryptic task and can be hard to democratize. On the other hand, even taking into account the no free lunch theorem, there are some techniques and process to automate the creation of optimized models. This article presents easyML a script for automation of features selection, features engineering and training of a model.
Recently, a fellow data enthusiast, https://firstname.lastname@example.org shared a tutorial on data visualization using the kaggle pokemon dataset. This made me wonder if there is a way to beat the game using some basic data analysis. I’ll present an analysis of the dataset for the version red and blue, which is the best starter and whether or not Ash actually stood a chance.
Deepfake consist in switching ones face in a video or picture for someones’ else face in the same picture in the best way possible. Advanced techniques based on deep learning and specifically auto-encoders along with the ease of access to computing power and image data enable the creation of very powerful models resulting in fake pictures that are hard to discriminate from real pictures. In this article, I’ll expose some of my experiments based on autoencoder for deepfake detection.
How to create your custom Optical Characters Recognition (OCR) system using python and computer vision. Based on the latest edition of Strange Adventures by DC comics, I realised that an allien language used there could be used for OCR.
Journalling about different Projects That I have been working on, how I implemented it and what I learned on the way