Word Sense Disambiguation On Supervised

Métriques

SemEval 2007

SemEval 2013

SemEval 2015

Senseval 2

Senseval 3

Résultats

Résultats de performance de divers modèles sur ce benchmark

						Paper Title
SANDWiCH	80.9	92.6	91.5	87.8	85.7	SANDWiCH: Semantical Analysis of Neighbours for Disambiguating Words in Context ad Hoc
ESR+WNGC	78.5	82.3	85.3	82.5	80.2	Improved Word Sense Disambiguation with Enhanced Sense Representations
ConSeC+WNGC	78.5	85.2	87.5	82.7	81.0	ConSeC: Word Sense Disambiguation as Continuous Sense Comprehension
ConSeC	77.4	83.2	85.2	82.3	79.9	ConSeC: Word Sense Disambiguation as Continuous Sense Comprehension
ESR	77.0	81.5	84.1	81.3	79.9	Improved Word Sense Disambiguation with Enhanced Sense Representations
ESCHER SemCor	76.3	82.2	83.2	81.7	77.8	ESC: Redesigning WSD with Extractive Sense Comprehension
EWISER+WNGC	75.2	80.7	81.8	80.8	79.0	Breaking Through the 80% Glass Ceiling: Raising the State of the Art in Word Sense Disambiguation by Incorporating Knowledge Graph Information
BEM	74.5	79.7	81.7	79.4	77.4	Moving Down the Long Tail of Word Sense Disambiguation with Gloss Informed Bi-encoders
SemCor+WNGC, hypernyms	73.4	78.7	82.6	79.7	77.8	Sense Vocabulary Compression through the Semantic Knowledge of WordNet for Neural Word Sense Disambiguation
SparseLMMS+WNGC	73.0	79.4	81.3	79.6	77.3	Sparsity Makes Sense: Word Sense Disambiguation Using Sparse Contextualized Word Representations
GlossBERT	72.5	76.1	80.4	77.7	75.2	GlossBERT: BERT for Word Sense Disambiguation with Gloss Knowledge
EWISER	71.0	78.9	79.3	78.9	78.4	Breaking Through the 80% Glass Ceiling: Raising the State of the Art in Word Sense Disambiguation by Incorporating Knowledge Graph Information
ARES	71.0	77.3	83.2	78.0	77.1	With More Contexts Comes Better Performance: Contextualized Sense Embeddings for All-Round Word Sense Disambiguation
SparseLMMS	68.8	76.1	77.5	77.9	77.8	Sparsity Makes Sense: Word Sense Disambiguation Using Sparse Contextualized Word Representations
BERT (linear projection)	68.1	71.1	76.2	75.5	73.6	Improved Word Sense Disambiguation Using Pre-Trained Contextualized Word Representations
SemCor+WNGT, vocabulary reduced, ensemble	66.81	72.63	74.46	75.15	70.11	Improving the Coverage and the Generalization Ability of Neural Word Sense Disambiguation through Hypernymy and Hyponymy Relationships
Bi-LSTM<sub>att+LEX+POS</sub>	64.8*	66.9	71.5	72.0	69.1	Neural Sequence Learning Models for Word Sense Disambiguation
Bi-LSTM<sub>att+LEX</sub>	63.7*	66.4	72.4	72.0	69.4	Neural Sequence Learning Models for Word Sense Disambiguation
BERT (nearest neighbour)	63.3	69.2	74.4	73.8	71.6	Improved Word Sense Disambiguation Using Pre-Trained Contextualized Word Representations
supWSD<sub>emb</sub>	63.1	66.8	71.8	72.7	70.6	-

0 of 26 row(s) selected.

Command Palette

Word Sense Disambiguation On Supervised

Métriques

Résultats