Dec 31 2017 - As part of a project I was working on, I had to read the research paper Skip-Thought Vectors by Kiros et. al with a magnifying glass and also implement it in PyTorch. In this process, I learnt quite a lot about why Skip-Thought works so well despite being very straightforward in principle. In this blog, I would like to share what goes behind the scenes in Skip-Thoughts — small things that actually make it work.
Aug 28 2017 - I recently came across an ICML’17 paper “Attentive Recurrent Comparators” which proposes a simple yet powerful model for data efficient learning. The paper presents the first super-human One-shot Classification performance on the Omniglot dataset using only raw pixel information! In this blog post I am going to present my understanding of the main ideas of the paper. The authors of the paper also released an implementation using Theano. But I found the implementation a bit difficult to follow. So as an exercise I actually went ahead and re-implemented it in PyTorch. Code at https://github.com/sanyam5/arc-pytorch