AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Jointly Extracting Multiple Triplets with Multilayer Translation Constraints

Zhen Tan, Xiang Zhao, Wei Wang, Weidong Xiao

AAAI 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add relation part (RP) to tri-part tagging (ablation)TME · NYT-multi (top-2)F1: 0.486 -> 0.567Improved
add type part (TP) to tagging scheme (ablation)TME · NYT-multi (top-2)F1: 0.561 -> 0.567Improved
add both TP and RP (full tri-part tagging) (ablation)TME · NYT-multi (top-2)F1: 0.470 -> 0.567Improved
add multi-layer translation (Bi-LSTM layer score functions) over embedding-only (ablation)TME · NYT-multi (top-2)F1: 0.433 -> 0.567Improved
add embedding-layer score function to Bi-LSTM-only translation (ablation)TME · NYT-multi (top-2)F1: 0.431 -> 0.567Improved
add character embedding (ablation)TME · NYT-multi (top-2)F1: 0.552 -> 0.567Improved
add CRF layer (ablation)TME · NYT-multi (top-2)F1: 0.530 -> 0.567Improved
add dropout (ablation)TME · NYT-multi (top-2)F1: 0.556 -> 0.567Improved
add pre-training (ablation)TME · NYT-multi (top-2)F1: 0.530 -> 0.567Improved
use same relation embedding across layers instead of extra embeddings (TME-MR) (ablation)TME · NYT-multiF1: 0.507 -> 0.551Improved
use learned relation embeddings instead of random stable ones (TME-RR) (ablation)TME · NYT-multiF1: 0.437 -> 0.551Improved
add negative sampling (vs TME-NS) (ablation)TME · NYT-multiF1: 0.525 -> 0.551Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy