Phase Conductor on Multi-layered Attentions for Machine Comprehension

ICLR 2018 Rui LiuWei WeiWeiguang MaoMaria Chikina

Attention models have been intensively studied to improve NLP tasks such as machine comprehension via both question-aware passage attention model and self-matching attention model. Our research proposes phase conductor (PhaseCond) for attention models in two meaningful ways... (read more)

PDF Abstract

Code


No code implementations yet. Submit your code now

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Question Answering SQuAD1.1 Conductor-net (ensemble) EM 76.996 # 83
F1 84.630 # 82
Question Answering SQuAD1.1 Conductor-net (single) EM 73.240 # 114
F1 81.933 # 109
Question Answering SQuAD1.1 Conductor-net (single model) EM 74.405 # 104
F1 82.742 # 102
Question Answering SQuAD1.1 dev PhaseCond (single) EM 72.1 # 20
F1 81.4 # 24

Methods used in the Paper