Multimodal Sentiment Analysis on CMU-MOSI Word Aligned (test) benchmark leaderboard · SOTA2 Research

Multimodal Sentiment Analysis on CMU-MOSI Word Aligned (test)

43.3Accuracy (7-Class)TupleInfoNCE

Evaluation Results

Method	Links
TupleInfoNCE2021.07		43.3	83.6	83.8	—	—
MISA2021.07		42.3	83.4	83.6	—	—
MMV FAC2021.07		41.5	83.5	83.4	—	—
InfoNCE2021.07		40.5	83.1	82.8	—	—
MulT2019.06		40	83	82.8	0.871	0.698
Train from scratch2021.07		40	83	82.8	—	—
Full modality (ours)2023.12		39.7	82.9	82.7	0.87	0.694
CMC2021.07		39.5	83.3	83	—	—
Ours2023.12		39.1	82.3	82.2	0.858	0.691
Language and vision (ours)2023.12		38.7	81.3	81.2	0.849	0.688
RMFN2019.06		38.3	78.4	78	0.922	0.681
RMFN2023.12		38.3	78.4	78	0.922	0.681
Language only (ours)2023.12		36.4	78.1	77.3	0.751	0.619
MFM2019.06		36.2	78.1	78.1	0.951	0.662
MFM2023.12		36.2	78.1	78.1	0.951	0.662
MCTN2019.06		35.6	79.3	79.1	0.909	0.676
MCTN2021.11		35.6	79.3	79.1	0.909	0.676
MCTN2023.12		35.6	79.3	79.1	0.909	0.676
LF-LSTM2019.06		35.3	76.8	76.7	1.015	0.625
LF-LSTM2021.11		35.3	76.8	76.7	1.015	0.625
MoNIG2021.11		34.1	80.6	80.6	0.951	0.68
Vision only (ours)2023.12		34.1	76.6	76.4	0.721	0.421
EF-LSTM2019.06		33.7	75.3	75.2	1.023	0.608
EF-LSTM2021.11		33.7	75.3	75.2	1.023	0.608
Evidential2021.11		33.4	77.6	77.6	0.974	0.655
RAVEN2019.06		33.2	78	76.6	0.915	0.691
RAVEN2021.11		33.2	78	76.6	0.915	0.691
RAVEN2023.12		33.2	78	76.6	0.915	0.691
MoNIG2021.11		33	80.2	80.4	0.959	0.664
Gaussian2021.11		32.9	78.4	78.4	0.982	0.657