Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Rethinking Self-Attention: Towards Interpretability in Neural Parsing

About

Attention mechanisms have improved the performance of NLP tasks while allowing models to remain explainable. Self-attention is currently widely used, however interpretability is difficult due to the numerous attention distributions. Recent work has shown that model representations can benefit from label-specific information, while facilitating interpretation of predictions. We introduce the Label Attention Layer: a new form of self-attention where attention heads represent labels. We test our novel layer by running constituency and dependency parsing experiments and show our new model obtains new state-of-the-art results for both tasks on both the Penn Treebank (PTB) and Chinese Treebank. Additionally, our model requires fewer self-attention layers compared to existing work. Finally, we find that the Label Attention heads learn relations between syntactic categories and show pathways to analyze errors.

Khalil Mrini, Franck Dernoncourt, Quan Tran, Trung Bui, Walter Chang, Ndapa Nakashole• 2019

Related benchmarks

TaskDatasetResultRank
Constituent ParsingPTB (test)
F196.38
127
Dependency ParsingChinese Treebank (CTB) (test)
UAS94.6
99
Dependency ParsingPenn Treebank (PTB) (test)
LAS96.3
80
Constituent ParsingCTB (test)
F1 Score92.64
45
Constituency ParsingCTB 5.1 (test)
F1 Score92.64
25
Dependency ParsingPTB
UAS97.4
24
Constituency ParsingCTB 5.0 (test)
F1 Score92.64
19
Constituency ParsingPTB (test)
Latency (s)40.8
3
Showing 8 of 8 rows

Other info

Code

Follow for update