@YiMaTweets A good theoretical paper not only tries to explain why the method works (which is usually not a complete explanation), but also proposes new methods/verifications according to the theory. I think that's essential for acceptance :)
Excited to receive the #SoCalNLP Best Paper Award for our paper "Empowering Language Models with Knowledge Graph Reasoning for Question Answering". The paper link is: https://t.co/7C6HzkebAK
Thanks to the organizers and all the great collaborators!
🎉 New preprint! Generate rather than Retrieve: Large Language Models are Strong Context Generators. Our proposed method achieved new SoTA on open-domain QA! (1/5)
Arxiv link: https://t.co/hfzN9bAneC
Our latest model KEAR surpasses human performance on the famous CommonsenseQA leaderboard. Paper: https://t.co/ixiZ9eBP9E Code: https://t.co/rqiNFCJGJi Video: https://t.co/4BMWasfr0x https://t.co/vTwc5n2rHO
Check out our latest paper on combining Dictionary+ConceptNet to achieve new SOTA on CommonsenseQA. Paper: https://t.co/WGSNqGbjYV Code: https://t.co/OUlECdYIN0
Growing up in Chicagoland in the 80's & 90's, I was subjected to many acts of racism.
I've been called chink and slanty-eyes. People made "ching-chong" noises around me. Someone once insisted on calling me Nissan because it was easier than remembering my actual name.
New post on ML@CMU blog about building binary classification and regression models that require fewer directly labeled examples by taking advantage of comparisons. Written by Yichong Xu and edited by @SimonShaoleiDu!
We investigate how to use pairwise comparisons instead of direct labels for machine learning. Check out the blog post on ML CMU blog:
https://t.co/dxDeVJm4f9
Check out our new paper about multi-task learning for machine reading comprehension:
https://t.co/mAahhGmemE
State-of-art results on NewsQA, Who-Did-What
Tomorrow @NAACLHLT, session 6D :)
#naacl2019