Djoerd Hiemstra @hiemstra@mastodon.utwente.nl

The annotated transformer: An annotated version of the "Attention is all you need" paper with code.

https://nlp.seas.harvard.edu/2018/04/03/attention.html

Xerte provides free software for educational content creation

https://xerte.org.uk/index.php/en/

Being adventurous and buying the @PINE64 phone, or maybe the pre-installed /e/OS Galaxy S7 refurbished to be on the save side? #FirstWorldProblems

https://e.foundation/product/e-os-galaxy-s7-refurbished/

I haven't checked my inbox since Christmas! Thank you InfoSupport!

Great read on independently reproducing machine learning papers: https://thegradient.pub/independently-reproducible-machine-learning/

#Vacancy We're hiring! (Assistant/Associate) Professor positions in Machine Learning for Data Science:

https://www.ru.nl/werken-bij/vacature/details-vacature/?recid=1090598

There is no data than more data seems to apply even to artificial data: Just generate training data from models like GPT. I wonder if we can apply this to (standard) search tasks too.

https://arxiv.org/abs/1911.03118

It turns out that Google is now using Chrome to track its users(!)

DNT: 1

https://www.magiclasso.co/insights/hey-google/

I signed the petition to allow remote paper & poster presentations at scientific conferences, in particular machine learning and data science conferences.

https://www.change.org/p/organizers-of-data-science-and-machine-learning-conferences-neurips-icml-aistats-iclr-uai-allow-remote-paper-poster-presentations-at-conferences

Congrats, @jantrienes for getting the best paper award at HSDM 2020 !

CLIN 30 keynote by Joakim Nivre : From Universal Dependencies to Sesame Street 🙂

https://clin30.sites.uu.nl/keynote/

We, computer Scientists should be less focused on conference publications... Good opinion piece by Moshe Vardi !

https://cacm.acm.org/magazines/2020/1/241717-publish-and-perish/fulltext

I did the first Data Science coffee talk in a very long time, discussing the usage of word piece models by BERT:

https://djoerdhiemstra.com/2020/some-thoughts-on-bert-and-word-piece-models/

More efficient and larger transformer models:

https://ai.googleblog.com/2020/01/reformer-efficient-transformer.html

Snorkel pogrammatically builds training data for ML: https://www.snorkel.org

The unstoppable rise of Sci-Hub:

https://blogs.lse.ac.uk/impactofsocialsciences/2019/02/19/the-unstoppable-rise-of-sci-hub-how-does-a-new-generation-of-researchers-perceive-sci-hub/

BERT word embedding tutorial by Chris McCormick and Nick Ryan:

http://mccormickml.com/2019/05/14/BERT-word-embeddings-tutorial/

ACM charges 15 dollar for reading Diane Kelly's one page #ICTIR2019 keynote abstract. It even charges me, a "professional member" 10 dollar... 😞

I signed this. The ACM's digital library should have gone full open access a long time ago!

https://www.change.org/p/association-for-computing-machinery-acm-support-open-access

Happy New Year!

https://www.youtube.com/watch?v=xobCI1gkK1M