The COVID-19 pandemic has caused an enormous amount of changes in how people work, play, and communicate. By now, many of us have settled into the routine of using remote communication or videoconferencing tools to keep in touch with our friends and family. In the last few weeks we've also seen a number of lists and guides aiming to get people set up with the "right" tools for communicating in hard times, but in almost every case, these articles recommend that people make a difficult compromise: trading their freedom in order to communicate with the people they care about and work with.

There are countless lists on the internet claiming to be the list of must-read programming books and it seemed that all those lists always recommended that same books minus two or three odd choices.

Finding good resources for learning programming is always tricky. Every-one has its own opinion about what book is the best to learn, and as we say in french, “Color and tastes should not be argued about”.

However I though it would be interesting to trust the wisdom of the crown and to find the books that appeared the most in those “Best Programming Book” lists.

If you want to jump right on the results go take a look below at the full results. If you want to learn about the methodology, bear with me.

Deep Learning has shown very promising results in the field of Computer Vision. But when applying it to practical domains such as medical imaging, lack of labeled data is a major challenge.

In practical settings, labeling data is a time consuming and expensive process. Though, you have a lot of images, only a small portion of them can be labeled due to resource constraints. In such settings, how can we leverage the remaining unlabeled images along with the labeled images to improve the performance of our model? The answer is semi-supervised learning.

FixMatch is a recent semi-supervised approach by Sohn et al. from Google Brain that improved the state of the art in semi-supervised learning(SSL). It is a simpler combination of previous methods such as UDA and ReMixMatch. In this post, we will understand the concept of FixMatch and also see how it got 78% median accuracy and 84% maximum accuracy on CIFAR-10 with just 10 labeled images.

Great final scene.