Skip to content

Latest commit

 

History

History
160 lines (127 loc) · 16.4 KB

Results-Simple.md

File metadata and controls

160 lines (127 loc) · 16.4 KB
Raw

The results below are subject to further verification and subdivision.

This is for reference only.



Results of IEMOCAP

Methods Modality paper Time WA/Accuracy(4) F1 Score(4) UA/Recall(4) Percision
M3ER VAT paper AAAI 2020 82.7 82.4 - -
LMFN VAT paper 2020 TMM 82.45 82.54 - -
HFFN VAT paper ACL 2019 82.37(4) 82.42(4) - -
MHA AT paper ICASSP 2019 76.5(WA-4) - 77.6(UA-4) -
HFusion VAT paper 2018 KBS 76.5 76.8 - -
CDSA VAT paper ACL 2017 76.1 - - -
LAMER AT paper InterSpeech 2019 70.4(4) - 69.5(4) -
MSER AT paper 2019 70.1(6) 71.8(6) 71.5(6) 72.9(6)
ASER A paper ICASSP 2017 63.5(WA-4) - 58.8(UA-4) -
MMResLSTM AT paper MM 2019 51.6(4_7) 50.3(4_7) - -


Methods Modality paper Time Acc(4) F1 Score(4) Recall(4) Percision Remark
M3ER VAT paper AAAI 2020 82.7 82.4 - - (4,ang,hap,neu,sad), (MA=Average(4), F1=Average(4))

Results of IEMOCAP

Methods Modality paper Time HAPPY-ACC HAPPY-F1 SAD-ACC SAD-F1 ANGRY-ACC ANGRY-F1 NEUTRAL-ACC NEUTRAL-F1
MulT VAT paper 2019 NIH Public Access 90.7/84.8 88.6/81.9 86.7/77.7 86.0/74.1 87.4/73.9 87.0/70.2 72.4/62.5 70.7/59.1
Multimodal Routing VAT paper EMNLP 2020 87.3 84.7 85.7 85.2 87.9 87.7 70.4 70.0
M3ER VAT paper AAAI 2020 81.6 0.862 88.1 0.828 86.8 0.862 74.4 0.745
HFusion VAT paper 2018 KBS 74.3 81.4 75.6 77.0 79.6 77.6 78.4 71.2
MTEE VAT paper AACL-IJCNLP 2020 85.0 - 86.6 - 88.1 - 71.1 -
CTNet AT paper 2021 TASLP 83.5 - 86.1 - 80.0 - 83.6 -
CDSA AVT paper ACL 2017 79.31 - 78.30 - 77.98 - 69.92 -
MMResLSTM AT paper MM 2019 61.0（h+e） - 74.1 - 31.8 - 54.1(n+f) -
HFFN VAT paper ACL 2019 - 88.65 - 86.24 - 94.31 - 76.24
HPFN VAT paper NeurIPS 2019 - 86.2 - 86.6 - 88.8 - 72.5
LRMF VAT paper ACL 2018 - 85.8 - 85.9 - 89.0 - 71.7



Results of CMU-MOSEI

Methods Modality paper Time Accuracy(B) F1 Score(B) Accuracy F1 Score MAE ABC
MSAF V paper 2020 85.5(2) 85.5(2) 52.4(7) - 0.559 0.738
MISA VAT paper ACM 2020 83.6/85.5(2) 83.8/85.3(2) 52.2(7) - 0.555 0.756
CM-BERT TA paper ACM MM 2020 83.5(2) 83.6(2) - - -
Self-MM VAT paper AAAI 2021 82.81/85.17(B) 82.53/85.30(B) 48.927(7) - 0.530 0.765
MulT VAT paper 2019 NIH Public Access 82.5/81.6 82.3/81.6 51.8/51.7(7) 0.580/0.591 0.703/0.694
Multimodal Routing VAT paper EMNLP 2020 81.7(B) 81.8(B) 51.6(7) - - -
Deep-HOSeq AVT paper ICDM 2020 74.32 $\pm$ 75.12 $\pm$ 44.17 $\pm$ 0.7189 $\pm$ 0.5438
SLAM [paper] ACL 2018 HML-5 69.4(B) 69.3(B) - -
MMRTN AT [paper] ACL 2018 HML-3 66.8(2) 0.63(2) 49.17(2) 0.45(2) 0.58 -
LMFN VAT paper 2020 TMM 61.31(T) 59.48(T) - -
SWAFN VAT paper ACL 2020 61.03(T) 59.32(T) - - - -
HFFN VAT paper ACL 2019 60.37(T) 59.07(T) - - - -
M3ER VAT paper AAAI 2020 - - 89.0(6) 90.2(6) - -
CAN [paper] ACL 2018 HML-4 - - 88.89(6) 0.89(6) - -
MTEE VAT paper AACL-IJCNLP 2020 - - 66.2(6) - -



Results of CMU-MOSI

Methods Modality paper Time Accuracy (B) F1 Score (B) Accuracy (7) MAE(R) Correlation(R)
Self-MM VAT paper AAAI 2021 84.00/85.98(B) 84.42/85.95(B) - 0.713 0.798
CM-BERT TA paper ACM MM 2020 84.5(2) 84.5(2) 44.9(7) 0.729 0.791
MulT VAT paper 2019 NIH Public Access 83.0/81.1 82.8/81.0 40.0/39.1(7) 0.871/0.889
MISA VAT paper ACM 2020 81.8/83.4(2) 81.7/83.6(2) 42.3(7) 0.783 0.761
MLMA VAT paper ICDM 2017 81.3(B) - - - -
LMFN VAT paper 2020 TMM 80.85(B) 80.92(B) - - -
CDSA VAT paper ACL 2017 80.3 - - - -
SWAFN VAT paper ACL 2020 80.2(B) 80.1(B) 40.1(7) 0.880 0.697
HFFN VAT paper ACL 2019 80.19(B) 80.34(B) - - -
HFusion VAT paper 2018 KBS 80.0(B) - - - -
HPFN VAT paper NeurIPS 2019 77.5(2) 77.4(2) 36.7(7) 0.945 0.672
TFN VAT paper EMNLP 2017 77.1(B) 77.9(B) 42.0(5) 0.87 0.70
LRMF VAT paper ACL 2018 76.4(2) 75.7(2) 32.8(7) 0.912 0.668
SLAM [paper] ACL 2018 HML-5 74.6(B) 74.1(B) - - -
DeepCU VAT paper IJCAI 2019 73.54 $\pm$ 1.10 73.52 $\pm$ 1.14 34.04 $\pm$ 3.61 1.0442 $\pm$ 1.71 $\times$ $10^{-5}$ 0.5609 $\pm$ 1.05 $\times$ $10^{-5}$
Deep-HOSeq AVT paper ICDM 2020 - - 35.87(7) 1.0201 0.5676



Results of MELD

Methods Modality paper Time Anger Disgust Fear Joy Neutral Sadness Surprise WAACC WAF1 UAF1
MMResLSTM AT paper MM 2019 79.4, 75.3 - - 70.4,70.1 65.7,65.4 84.0,79.2 78.3,74.0 - - -
CTNet-4 AT paper 2021 TASLP 44.6 11.2 10.0 56.0 77.4 32.5 52.7 62.0 60.5 40.6



Results of MOUD

Methods Modality paper Time Accuracy F1 Score MSE ABC
CDSA VAT paper ACL 2017 68.11(B) - - -



Results of POM

Methods Modality paper Time Accuracy F1 Score MAE Correlation
LRMF VAT paper ACL 2018 42.8 - 0.796 0.396
DeepCU VAT paper IJCAI 2019 34.77 $\pm$ 0.493 - 0.8568 $\pm$ 0.0045 0.2845 $\pm$ 0.0009



Results of RAVDESS

Methods Modality paper Time Accuracy F1 Score MAE Correlation
MSAF V paper 2020 74.86 - - -



Methods paper Time Coding Python Pytorch Tensorflow Keras
Self-MM paper AAAI 2021 coding Python 3.7 torch 1.2.0 - -
CTNet paper 2021 TASLP -
SWAFN paper ACL 2020 coding1 coding2 Python 3.6 pytorch 1.4.0
MISA paper ACM 2020 coding Python 3.7.5 Pytorch 1.3.1
CM-BERT paper MM 2020 coding Python 3.x torch 1.2.0
Multimodal Routing paper EMLP 2020 coding Python 3.6 Pytorch (>=1.2.0)
SF-SSL paper InterSpeech 2020 coding Python 3 Pytorch
MTEE paper AACL-IJCNLP 2020 coding Python 3.6 + PyTorch 1.4 +
SSE-FT paper 2020 Access coding Python Pytorch
MSAF paper 2020 coding Python 3.6.12 Pytorch 1.7
HPFN paper NeurIPS 2019 coding Python 3 Pytorch
MulT paper 2019 NIH Public Access coding Python 3 Pytorch
MSER paper 2019 coding Python 3.6/3.7 Pytorch (>=1.0.0)
LRMF paper ACL 2018 coding Python 2.7/3.6+ torch 0.3.1
TFN paper EMNLP 2017 coding Python PyTorch
ASER paper ICASSP 2017 coding Python 3.6.4 Pytorch 0.4.1
Deep-HOSeq paper ICDM 2020 coding Python 3.6 TensorFlow 1.12.0
LMFN paper 2020 TMM coding python 2.x - Keras 2.0.6
HFFN paper ACL 2019 coding python 2.x - Keras 2.0.6
MMResLSTM paper MM2019 coding Python 3.6 TensorFlow 1.3.0 Keras 2.0.9
DEEPCU paper IJCAI 2019 coding Python 3.6 TensorFlow 1.12.0
MHA paper ICASSP 2019 coding Python 3.7 tensorflow 1.14
LAMER paper InterSpeech 2019 coding Python 3.6 TensorFlow 2.x
MRTN paper ACL-Challenge-HML 2018 coding Python 3.6 TensorFlow 2.x
HFusion paper 2018 KBS coding python 3.x Tensorflow >= 1.7 Keras >= 2.0
CDSA paper ACL 2017 coding Python tensorflow 1.x
MLMA paper ICDM 2017 coding Python Tensorflow 1.x
EEMER paper 2017 IJSTSP coding Python <= 2.7 TensorFlow <= 0.12
PI [paper] ACL-WK 2018 coding -