Skip to main content

Extrinsic task performance leaderboard (all vector size data)

Extrinsic task performance leaderboard (all vector size data)

See LDT paper for implementation details on all tasks.

Extrinsic task datasets:

  • POS-tag.: POS-tagging task from CoNLL 2003 dataset (Tjong Kim Sang and De Meulder, 2003);
  • Chunk.: Chunking task from CoNLL 2003 dataset (Tjong Kim Sang and De Meulder, 2003);
  • NER: Named Entity Recognition task from CoNLL 2003 dataset (Tjong Kim Sang and De Meulder, 2003);
  • Rel.Clas.: semantic relation classification task from SemEval 2010 (Hendrickx et al., 2010);
  • Subj. Clas.: Rotten Tomato dataset (Pang and Lee, 2004);
  • MR: short movie review dataset for sentence-level sentiment classification (Pang and Lee, 2005);
  • IMDB: Standofd IMDB movie review dataset (Maas et al., 2011);
  • SNLI: Stanford Natural Language Inference dataset (Bowman et al., 2015).
Model Context Dims. POS-tag. Chunk. NER Rel. Clas. Subj. Clas. MR IMDB SNLI
CBOW structured_deps 100 88.89 77.94 76.58 71.69 88.04 73.31 78.01 64.34
CBOW structured_deps 25 87.68 72.1 71.28 68.41 82.88 68.29 67.74 54.46
CBOW structured_deps 250 88.51 78.05 76.09 72.23 88.52 73.91 81.24 67.79
CBOW structured_deps 50 88.61 74.58 74.95 69.99 85.24 70.93 74.22 60.44
CBOW structured_deps 500 88.34 78.03 75 73.92 87.36 73.88 82.11 69.33
GloVe structured_deps 100 87.65 74.86 72.17 71.47 86.36 73.59 79.99 64.02
GloVe structured_deps 25 84.75 66.43 64.71 68.39 81.96 66.54 70.72 53.52
GloVe structured_deps 250 87.61 75.77 72.1 72.95 87.24 74.15 83.03 67.48
GloVe structured_deps 50 86.76 72.31 69.58 69.26 84 69.8 74.97 59.55
GloVe structured_deps 500 86.98 75.76 70.05 72.83 87.76 74.8 76.08 68.82
SkipGram structured_deps 100 88.9 78.35 76.61 71.39 88.26 75.51 78.39 64.66
SkipGram structured_deps 25 87.81 72.37 69.85 69.76 82.96 69.97 68.07 53.87
SkipGram structured_deps 250 88.57 78.88 75.94 73.76 89 75.35 81.07 68.21
SkipGram structured_deps 50 88.71 74.05 74.05 71.14 87.24 72.81 74.62 59.64
SkipGram structured_deps 500 88.44 78.82 75.86 74.2 89.56 75.42 81.71 69.87
CBOW structured_linear 100 88.52 78.6 75.48 71.46 88.2 73.41 77.54 63.79
CBOW structured_linear 25 86.31 71.25 71.64 66.69 84.44 68.64 63.66 51.28
CBOW structured_linear 250 88.63 79.2 75.25 73.88 88.76 74.86 82.15 67.7
CBOW structured_linear 50 87.96 75.68 75.38 71.14 86.92 71.99 72.35 58.6
CBOW structured_linear 500 88.12 78.29 75.24 74.48 88.96 75.12 82.66 69.19
GloVe structured_linear 100 88.09 77.2 75.1 72.12 87.52 72.81 77.77 63.99
GloVe structured_linear 25 85.55 69.85 68.31 68.28 84.52 68.02 63.62 54.12
GloVe structured_linear 250 88.16 78.09 74.69 72.32 88.44 74.09 81.38 67.66
GloVe structured_linear 50 87.44 74.45 74.1 69.67 86.2 71.41 71.56 59.33
GloVe structured_linear 500 87.67 77.45 74.12 72.57 89.28 75.59 82.68 69.11
SkipGram structured_linear 100 88.56 78.49 76.45 73.84 88.84 74.56 78.75 64.28
SkipGram structured_linear 25 86.31 71.29 72.61 67.34 85.2 70 66.36 53.65
SkipGram structured_linear 250 88.58 79.15 75.79 74.7 89.32 75.82 81.67 68.37
SkipGram structured_linear 50 87.98 74.96 75.7 71.38 86.68 74.18 73.83 60.19
SkipGram structured_linear 500 88.31 79.14 75.54 75.6 90.2 76.91 82.95 69.42
CBOW word_deps 100 87.56 75.8 73.59 70.41 88.56 73.24 75.62 63.4
CBOW word_deps 25 82.17 63.36 68.82 66.08 85.16 70.33 61.05 53.03
CBOW word_deps 250 88.06 77.84 76.2 72.62 89.6 74.15 79.61 67.56
CBOW word_deps 50 86.4 71.55 73.32 69.93 86.52 72.73 69.72 58.26
CBOW word_deps 500 87.71 77.56 73.58 73.53 89.7 75.46 81.25 69.14
GloVe word_deps 100 83.1 62.79 68.74 70.43 88 72.69 78.14 64.51
GloVe word_deps 25 75.96 51.3 63.37 66.95 85.96 68.28 69.49 55
GloVe word_deps 250 84.34 66.86 69.2 70.47 88.16 74.31 81.13 67.3
GloVe word_deps 50 80.75 58.08 66.19 68.82 86.92 70.49 73.01 59.42
GloVe word_deps 500 84.67 68.36 68.76 72.02 88.8 74.69 82.57 69.28
SkipGram word_deps 100 87.62 75.29 75.63 72.28 89.2 74.5 76.95 64
SkipGram word_deps 25 82.37 62.58 69.19 66.52 85.68 70.66 68.27 54.24
SkipGram word_deps 250 88.17 77.23 76.59 73.8 89.84 75.66 81.38 68.09
SkipGram word_deps 50 86.44 69.06 73.48 71.2 87.48 73.85 71.41 59.56
SkipGram word_deps 500 87.77 78.05 75.38 73.96 89.8 77.15 82.78 69.93
CBOW word_linear 100 87.8 76.51 75.65 71.51 88.56 74.95 76.7 63.81
CBOW word_linear 25 81.91 66.98 70.14 65.61 85.76 70.44 60.2 49.71
CBOW word_linear 250 88.28 78.52 75.33 74.64 89.76 75.09 81.46 67.66
CBOW word_linear 50 86.37 71.86 73.73 69.51 87.68 72.84 69.06 59.13
CBOW word_linear 500 87.66 77.53 75.21 74.78 89.8 75.9 82.22 69.29
GloVe word_linear 100 82.15 61.18 69.16 68.03 88.36 72.6 78.62 63.93
GloVe word_linear 25 76.25 51.1 64.24 64.9 85.96 66.97 68.67 54.07
GloVe word_linear 250 83.72 64.64 69.5 70.8 89.2 73.46 81.15 67.79
GloVe word_linear 50 80.22 57.28 65.97 67.82 87.6 70.84 73.25 59.87
GloVe word_linear 500 83.8 66.1 69.62 71.05 89.16 74.6 82.24 69.51
SkipGram word_linear 100 87.91 76.46 76.75 72.83 89.12 73.79 77.27 64.17
SkipGram word_linear 25 81.64 63.76 70.37 65.42 85.36 71.06 67.32 52.72
SkipGram word_linear 250 88.34 78.06 75.3 74.77 89.68 76.05 81.12 68.24
SkipGram word_linear 50 86.41 71.64 74.62 70.05 87.32 73.61 71.58 59.46
SkipGram word_linear 500 87.86 78.23 75.72 74.8 89.92 76.86 82.73 69.74