List your company
ResearchFormerly WizWand

Entity Disambiguation on AIDA-CONLL (test)

97.1In-KB AccuracyLUKE
87.53290.01692.594.984Sep 1, 2019Oct 28, 2019Dec 25, 2019Feb 21, 2020Apr 18, 2020Jun 15, 2020Aug 12, 2020
Updated 8mo ago

Evaluation results

MethodLinks
LUKEfine-tuning=true, decoding strategy=confidence-order, candidate set=PPRforNED2019.0997.1
LUKEfine-tuning=true, decoding strategy=natural-order, candidate set=PPRforNED2019.0997
LUKEfine-tuning=true, decoding strategy=local, candidate set=PPRforNED2019.0996.8
Févry et al.candidate set=PPRforNED2019.0996.7
Yang et al.candidate set=PPRforNED2019.0995.9
LUKEfine-tuning=true, decoding strategy=confidence-order, candidate set=KB+YAGO2019.0995
DCA-SL + Triplescontext=1-hop KG triples2020.0894.94
Ling et al.candidate set=PPRforNED2019.0994.9
DeepType (2018)2020.0894.88
LUKEfine-tuning=true, decoding strategy=natural-order, candidate set=KB+YAGO2019.0994.8
DCA-SL (2019)2020.0894.64
Yang et al.candidate set=KB+YAGO2019.0994.6
LUKEfine-tuning=false, decoding strategy=confidence-order, candidate set=PPRforNED2019.0994.6
LUKEfine-tuning=true, decoding strategy=local, candidate set=KB+YAGO2019.0994.5
Fang et al. (2019)2020.0894.3
Fang et al.candidate set=KB+YAGO2019.0994.3
LUKEfine-tuning=false, decoding strategy=natural-order, candidate set=PPRforNED2019.0994
LUKEfine-tuning=false, decoding strategy=local, candidate set=PPRforNED2019.0994
Chen et al. (2020)2020.0893.54
Cao et al.candidate set=KB+YAGO2019.0993.3
Yamada et al.candidate set=PPRforNED2019.0993.1
Le and Titovcandidate set=KB+YAGO2019.0993.1
Le & Titov (2018)2020.0893.07
Yang et al. (2018)2020.0893
Yang et al.candidate set=KB+YAGO2019.0993
Barba et al.candidate set=KB+YAGO2019.0992.6
Févry et al.candidate set=KB+YAGO2019.0992.5
LUKEfine-tuning=false, decoding strategy=confidence-order, candidate set=KB+YAGO2019.0992.4
Ganea & Hofmann (2017)2020.0892.22
Ganea and Hofmanncandidate set=KB+YAGO2019.0992.2
LUKEfine-tuning=false, decoding strategy=natural-order, candidate set=KB+YAGO2019.0991.7
Yamada et al. (2016)2020.0891.5
Yamada et al.candidate set=KB+YAGO2019.0991.5
LUKEfine-tuning=false, decoding strategy=local, candidate set=KB+YAGO2019.0990.8
Le & Titov (2019)2020.0889.66
Broscheitcandidate set=KB+YAGO2019.0987.9