In this post we are going to introduce a useful tool called Docker containers. I came across these recently when I decided to finally move my deep learning model off Google Colab and onto a computer equipped with GPUs.*1 The Biorobotics lab uses Docker containers so that every lab member…

I have been meaning to write this post about the Gumbel-softmax distribution for several months, but I put it on a back burner after I had dug myself into a hole of deep confusion and couldn’t get out. After some encouragement from my advisor, I decided to pick it up…

In my previous post, I presented the motivation for graph networks as described by a paper by DeepMind that I have been reading [1]. We have covered the ideas of relational inductive bias and combinatorial generalization, which describe how we, as humans, think. The DeepMind paper argues that using graph…

In my research I have been learning how neural networks can be used to process graphs as inputs, instead of scalar or vector values. This is a rich area of research and I thought I would start by reading an excellent review/position paper by a group of researchers at DeepMind…

This is the third post in a series on graph networks, based on a recent paper by researchers at DeepMind [1]. So far we have discussed the motivation for why graph networks are useful for advancing AI research, and how they work from a probability perspective. …

This week I am learning about the fundamentals of designing and building DNA origami structures. I should probably write a post just about cool stuff you can build with DNA origami to get you excited, but in the meantime, check out this cool TED Talk by one of the fathers…