papers

These are my notes from research papers I read. Each page’s title is also a link to the abstract or PDF.

It's not just size that matters: small language models are also few-shot learners

We presented this paper as a mini-lecture in Bang Liu’s IFT6289 course in winter 2022. You can view the slides we used here.

Scaling laws for transfer

This post was created as an assignment in Irina Rish’s neural scaling laws course (IFT6167) in winter 2022. The post contains no summarization, only questions and thoughts. Sometimes these scaling laws can feel like pseudoscience because they’re a post hoc attempt to place a trend line on data. How can we be confident that the trends we observe actually reflect the scaling laws that we’re after? In the limitations section they mention that they didn’t tune hyperparameters for fine-tuning or for the code data distribution.
Read more

Deep learning scaling is predictable, empirically

This was a paper we presented about in Irina Rish’s neural scaling laws course (IFT6167) in winter 2022. You can view the slides we used here. It’s important to note that in the results for NMT (Figure 1) we would expect the lines in the graph on the left to curve as the capacity of the individual models is exhausted. That’s why the authors fit the curves with an extra constant added.
Read more

Masked autoencoders are scalable vision learners

This post was created as an assignment in Irina Rish’s neural scaling laws course (IFT6167) in winter 2022. The post contains no summarization, only questions and thoughts. In this paper they mention that the mask vector is learned, and it sounds like the positional embeddings are also learned. I remember in Attention is all you need they found that cosine positional embeddings worked better than learned ones, especially for sequences of longer length.
Read more

Data scaling laws in NMT: the effect of noise and architecture

This paper is all about trying a bunch of different changes to the training setup to see what affects the power law exponent over dataset size. Here are some of the answers: encoder-decoder size asymmetry: exponent not affected, but effective model capacity affected architecture (LSTM vs. Transformer): exponent not affected, but effective model capacity affected dataset quality (filtered vs. not): exponent and effective model capacity not effected, losses on smaller datasets affected dataset source (ParaCrawl vs.
Read more