Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering

Open paper

Benchmark	Model	Rank	Results
question-answering-on-natural-questions	FiD-KD (full)	#5	EM: 54.7
question-answering-on-natural-questions	FID (full)	#8	EM: 51.4
question-answering-on-triviaqa	Fusion-in-Decoder (large)	#20	EM: 67.6