A paper that expands on the common methodology to compress high-dimensional datesets like embeddings into codebooks. They use this new technique called residual quantization which fixes the codebook's to a certain size and recursively quantizes the subsequent feature map in a coarse-to-fine manner.

A cool paper I found out at random. But a very cool application of NLP to make a audio-audio model. From what I understand this is a foundational model for LLMs which can understand and generate speech. In terms of contemporary research this is a bit old (2023), but foundational nonetheless!

A very interesting paper on the applications of Automatic Differentiation outside the domain of machine learning. It is a bit of a hard read, specially for someone a little out of touch on the CDF community. Nevertheless, it is a great read!

The paper explores the topic of Automatic/Algorithmic Differentiation (AD); an essential technology in contemporary machine learning models. It surveys a variety of methodologies to accomplish such concept and provides useful examples.

This revolutionary article pioneers the idea of Attention. One of the many algorithms utilized in the now renowned GPT LLMs OpenAI implements. Great read to understand how a transformer (the 'T' on 'GPT') functions.

A total existential book from what I have heard. Still reading

A beautiful book about the search for happiness. Depicting a 'classic' rags to riches story of a single-dad broker into a enterpreneur. Discussions like family trauma and crisis are unforgettable. Must read!