Trajectory Prediction on UNIV ETH-UCY (test)
0.21ADESocialVAE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SocialVAENumber of samples (S)=202024.03 | 0.21 | 0.36 | — | |
| MGFM=202024.02 | 0.21 | 0.39 | — | |
| NPSNNumber of samples (S)=202024.03 | 0.23 | 0.39 | — | |
| EqMotionNumber of samples (S)=202024.03 | 0.23 | 0.43 | — | |
| EqMotionM=202024.02 | 0.23 | 0.43 | — | |
| GP-GraphNumber of samples (S)=202024.03 | 0.24 | 0.42 | — | |
| EigenTrajectoryNumber of samples (S)=202024.03 | 0.24 | 0.43 | — | |
| AgentFormerNumber of samples (S)=202024.03 | 0.25 | 0.45 | — | |
| SingularTrajectoryNumber of samples (S)=202024.03 | 0.25 | 0.44 | — | |
| AgentFormerM=202024.02 | 0.25 | 0.45 | — | |
| GroupNetNumber of samples (S)=202024.03 | 0.26 | 0.49 | — | |
| LEDNumber of samples (S)=202024.03 | 0.26 | 0.43 | — | |
| GroupNetM=202024.02 | 0.26 | 0.49 | — | |
| LB-EBMProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.27 | 0.52 | — | |
| LBEBMNumber of samples (S)=202024.03 | 0.28 | 0.54 | — | |
| MIDNumber of samples (S)=202024.03 | 0.29 | 0.55 | — | |
| FlowChainM=202024.02 | 0.29 | 0.54 | — | |
| Trajectron++Number of samples (S)=202024.03 | 0.3 | 0.55 | — | |
| Trajectron++M=202024.02 | 0.3 | 0.55 | — | |
| MIDM=202024.02 | 0.3 | 0.55 | — | |
| PECNetNumber of samples (S)=202024.03 | 0.34 | 0.56 | — | |
| PECNetProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.35 | 0.6 | — | |
| STARNumber of samples (S)=202024.03 | 0.35 | 0.75 | — | |
| PECNetM=202024.02 | 0.35 | 0.6 | — | |
| Social-STGCNNProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.44 | 0.79 | — | |
| STGCNNNumber of samples (S)=202024.03 | 0.44 | 0.79 | — | |
| Social-STGCNNM=202024.02 | 0.44 | 0.79 | — | |
| CGNSProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.48 | 1.22 | — | |
| STSGNProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.48 | 1.08 | — | |
| SR-LSTM-2Probabilistic=false, History frames=8, Prediction frames=122021.04 | 0.51 | 1.1 | — | |
| STGATM=202024.02 | 0.52 | 1.1 | — | |
| SoPhieProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.54 | 1.24 | — | |
| Social-BiGATProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.55 | 1.32 | — | |
| GATProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.57 | 1.29 | — | |
| MATFProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.6 | 1.26 | — | |
| PIFProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.6 | 1.27 | — | |
| S-LSTMProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.67 | 1.4 | — | |
| S-GAN-PProbabilistic=true, History frames=8, Prediction frames=122021.04 | 0.76 | 1.52 | — | |
| Social-GANNumber of samples (S)=202024.03 | 0.76 | 1.52 | — | |
| Social-GANM=202024.02 | 0.76 | 1.52 | — | |
| LinearProbabilistic=false, History frames=8, Prediction frames=122021.04 | 0.82 | 1.59 | — | |
| Tunable Soft EquivarianceArchitecture=Autoregressive Transformer2026.03 | 7.85 | 8.07 | 0.69 | |
| BaseArchitecture=Autoregressive Transformer2026.03 | 7.91 | 8.16 | 0.73 | |
| EqAutoArchitecture=Equivariant Autoregressive Transformer2026.03 | 8.16 | 8.33 | 0 |