SocialIQA is the first large-scale benchmark with 38k crowdsourced questions testing commonsense about social interactions, where pretrained language models trail humans by over 20% but transfer to improve performance on Winograd Schemas and COPA.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CL 3years
2019 3verdicts
UNVERDICTED 3representative citing papers
Gated fusion of fastText and BERT embeddings into an end-to-end ASR model captures multi-sentence conversational context and lowers word error rate on the Switchboard corpus.
DAL combines dual learning on query-response pairs with adversarial training to improve diversity and naturalness in generated dialogue responses over prior methods.
citing papers explorer
-
SocialIQA: Commonsense Reasoning about Social Interactions
SocialIQA is the first large-scale benchmark with 38k crowdsourced questions testing commonsense about social interactions, where pretrained language models trail humans by over 20% but transfer to improve performance on Winograd Schemas and COPA.
-
Gated Embeddings in End-to-End Speech Recognition for Conversational-Context Fusion
Gated fusion of fastText and BERT embeddings into an end-to-end ASR model captures multi-sentence conversational context and lowers word error rate on the Switchboard corpus.
-
DAL: Dual Adversarial Learning for Dialogue Generation
DAL combines dual learning on query-response pairs with adversarial training to improve diversity and naturalness in generated dialogue responses over prior methods.