Table 3 The median of best test accuracy of three QNLP models on data of different length intervals.
From: Recurrent quantum embedding neural network and its application in vulnerability detection
Model | DisCoCat (%) | QLSTM (%) | RQENN (%) |
|---|---|---|---|
\(T=10(\Delta r=10)\) | 95.0 | 95.0 | 100 |
\(T=20(\Delta r=10)\) | 98.0 | 90.0 | 98.0 |
\(T=50(\Delta r=10)\) | – | 95.0 | 99.0 |
\(T=60(\Delta r=10)\) | – | 74.0 | 92.0 |
\(T=70(\Delta r=10)\) | – | 68.5 | 88.0 |
\(T=80(\Delta r=10)\) | – | 85.5 | 90.0 |
\(T=90(\Delta r=10)\) | – | 79.0 | 89.5 |
\(T=100(\Delta r=10)\) | – | 76.0 | 88.5 |
\(T=100(\Delta r=60)\) | – | 71.7 | 87.4 |
