Here’s all my notes from NAACL-HLT 2019!
Not present are all the fascinating conversations I had with random people, mostly over breakfast. 
(I don’t know what it is about breakfast! My theory is that it’s because everyone is undercaffeinated and hence both more humble about their own work and less self-conscious about asking dumb questions, but someone should seriously do a study on this.)
Sunday (tutorials)
Monday
- Arvind Narayanan on how sociology can illuminate models and ML can illuminate society
- Cognitive Track on using neuroscience to study neural models
Tuesday
- Industry Track on Alexa (seriously), and IE from dialogues and visual documents
- Semantics Track, which in this case seemed like a bit of all catch-all (I guess semantics is the bit of language that isn’t the other stuff)
- Rada Mihalcea on deception detection
Wednesday
- Kieran Snyder on data ethics and what a product is
- Question Answering Track on new QA datasets and some very complex models