[Adapt] Seminar]Annotation Artifacts in Natural Language Inference Data

[黄姗姗]

Hi Adapters,

Nowadays, full network pre-training, like BERT,  has led to a series of breakthroughs in language representation learning. Especially for some reading comprehension tasks, the deep models perform better than human. Does the machine already surpassed human with reasoning ability or only get the tricks in multiple choice questions? 
In this seminar, I will introduce some research about annotation artifacts in datasets. 
 		 	 	 		

Related papers:
Probing Neural Network Comprehension of Natural Language Arguments
Annotation Artifacts in Natural Language Inference Data 
Tackling the Story Ending Biases in The Story Cloze Test 

Time: Wed 4:30pm

Venue: SEIEE 3-414

Best regards,
Shanshan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.sjtu.edu.cn/pipermail/adapt/attachments/20191112/b191894b/attachment.html>