@kishore__vasan And when they inevitably go out battery in 30 minutes, they will tell you “I need a break from talking to you and recharge my social battery”
Really excited to share new work on the whose language is more likely to be represented in pretraining corpora for large language models, with @ssgrn@skdreier24 @ekgade Leroy Wang @blarrywang@LukeZettlemoyer and @nlpnoah!
Paper: https://t.co/EhuyA9iqk8
Added new features to the app with more data resolution!
It is great to see clinical trials already in the recruiting phase at some of the epicenters of the outbreak. 3 in Seattle and 16 in Wuhan! #COVID19#CoronaVirusUpdate
As we begin to contain this outbreak, conducting clinical trials are critical to coming up with a solution. I hope this application will provide some key insights into the progress in that space! #COVID19#CoronaVirusUpdates https://t.co/SovqXkc8eX
Here's one: "But advances in technology mean it is now possible to ask an artificial intelligence (AI) for its views on the coming year."
Sorry, a language model doesn't have "views". (Also, the answers turn out to have been cherry-picked.)
https://t.co/kphGFYvutC
Introducing the SHA-RNN :)
- Read alternative history as a research genre
- Learn of the terrifying tokenization attack that leaves language models perplexed
- Get near SotA results on enwik8 in hours on a lone GPU
No Sesame Street or Transformers allowed.
https://t.co/oCArjFKVDK
Dear ML and #NLProc researchers: Just because you can cast something as a seq2seq problem DOES NOT MEAN you should.
I'll go further: If you're ever tempted to do something like this, you need to take three giant steps back and ask yourself "why"? >>