In this episode of Machine Learning Street Talk, Tim Scarfe, Yannic Kilcher and Connor Shorten chat about Large-scale Transfer Learning in Natural Language Processing. The Text-to-Text Transfer Transformer (T5) model from Google AI does an exhaustive survey of what’s important for Transfer Learning in NLP and what’s not. In this conversation, we go through the key takeaways of the paper, text-to-text input/output format, architecture choice, dataset size and composition, fine-tuning strategy, and how to best use more computation.

Beginning with these topics, we diverge into exciting ideas such as embodied cognition, meta-learning, and the measure of intelligence. We are still beginning our podcast journey and really appreciate any feedback from our listeners. Is the chat too technical? Do you prefer group discussions, interviewing experts, or chats between the three of us? Thanks for watching and if you haven’t already, Please Subscribe!

Paper Links discussed in the chat:

Text-to-Text Transfer Transformer: https://arxiv.org/abs/1910.10683

Experience Grounds Language (relevant to divergent discussion about embodied cognition): https://arxiv.org/pdf/2004.10151.pdf

On the Measure of Intelligence: https://arxiv.org/abs/1911.01547

Train Large, Then Compress: https://arxiv.org/pdf/2002.11794.pdf

Scaling Laws for Neural Language Models: https://arxiv.org/pdf/2001.08361.pdf

The Illustrated Transformer: http://jalammar.github.io/illustrated...

ELECTRA: https://arxiv.org/pdf/2003.10555.pdf

Transformer-XL: https://arxiv.org/pdf/1901.02860.pdf

Reformer: The Efficient Transformer: https://openreview.net/pdf?id=rkgNKkHtvB

The Evolved Transformer: https://arxiv.org/pdf/1901.11117.pdf

DistilBERT: https://arxiv.org/pdf/1910.01108.pdf

How to generate text (HIGHLY RECOMMEND): https://huggingface.co/blog/how-to-ge...

Tokenizers: https://blog.floydhub.com/tokenization-nlp/