Thoughts on Kathy McKeown's #acl2020nlp keynote:
someone should go back and categorize all the "what's most important about deep learning" responses---literally nobody agrees! Accessibility, scalability, empirical performance, representations of lexical meaning, feature representations more broadly, support for new inputs, ...
Always good to be reminded that apart from Aravind, almost all of the "founding members" of what's now the NLP community were women: Spärck-Jones, Webber, Grosz, Hajicova.
This talk is a great summary of the various ways in which domain expertise will continue to be relevant even if we wind up in a world in which all modeling questions are boring.
Regina Barzilay: "NLP is very far from being like electricity. Sometimes you plug your coffee maker in and it turns into a blow dryer."
Strongly agree with the observation that better generation requires an explicit representation of communicative intent (check out my ConvAI workshop talk! slideslive.com/38929625/conve…); would have been cool get some perspectives on how this might be done non-symbolically.
And nice to end with a statement about research culture: disciplinary expertise is as valuable as ever in questions about the design of tasks and evaluations, and the first step towards getting these right is incentivizing people to take them seriously.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
