MUSTARD
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MUSTARD
95.3S-TEDS
20
MUStARD original (speaker-dependent)
86.8Precision
15
MUStARD speaker-independent original
75.6F1 Score
13
MUStARD
76.62Accuracy
13
MUSTARD
69.86Accuracy
13
MUSTARD++ (test)
83.2F1 Score
13
MUSTARD (test)
72.5Accuracy
12
MuSTARD++
83.2F1 Score
10
MUStARD Synergy
52.9Accuracy
8
MUStARD (held-out)
65.8F1 Score
8
MUStARD
0.795Weighted F1
7
MuSTARD
86.8Precision
7
MUStARD
80.6Accuracy
6
MUStARD (speaker dependent)
38.4ROUGE-1
6
MUStARD speaker dependent
86.8Precision
6
MUSTARD (test)
78.7F1 Score
5
MUStARD
81Recall@10 (Q: M23, T: M1)
4
MUStARD (5-fold CV)
74.42Accuracy
2
MUStARD++ (5-fold CV)
73.29Accuracy
1
MUSTARD (unseen)
95.15Performance
1