Human Activity Recognition on HAR
7.53Time per Sweep (s)AE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| AEModel Type=Auto-encoders, Feature Dimension=32, # of Params.=62,9402021.06 | 7.53 | — | — | |
| SALSModel Type=Tensor models, Rank=32, # of Params.=2,6882021.06 | 7.535 | — | 0 | |
| AEModel Type=Auto-encoders, Feature Dimension=128, # of Params.=87,6122021.06 | 7.662 | — | — | |
| AEssModel Type=Auto-encoders, Feature Dimension=32, # of Params.=62,9402021.06 | 7.765 | — | — | |
| AEssModel Type=Auto-encoders, Feature Dimension=128, # of Params.=87,6122021.06 | 7.804 | — | — | |
| GR-SALSModel Type=Tensor models, Rank=32, # of Params.=2,6882021.06 | 7.829 | — | 0.0003 | |
| BYOLModel Type=Self-sup models, Feature Dimension=32, # of Params.=54,3422021.06 | 8.43 | — | — | |
| SimCLRModel Type=Self-sup models, Feature Dimension=32, # of Params.=53,2862021.06 | 8.459 | — | — | |
| BYOLModel Type=Self-sup models, Feature Dimension=128, # of Params.=82,1822021.06 | 8.478 | — | — | |
| SimCLRModel Type=Self-sup models, Feature Dimension=128, # of Params.=65,6702021.06 | 8.532 | — | — | |
| ATDssModel Type=Tensor models, Rank=32, # of Params.=2,6882021.06 | 8.604 | — | 0.0011 | |
| ATDModel Type=Tensor models, Rank=32, # of Params.=2,6882021.06 | 8.672 | — | — | |
| Best Single Expert2026.03 | — | 91.2 | — | |
| EARCP2026.03 | — | 94.8 | — | |
| Equal Weighting2026.03 | — | 92.8 | — | |
| Hedge2026.03 | — | 93.9 | — | |
| Offline MoE2026.03 | — | 93.5 | — | |
| Stacking2026.03 | — | 93.1 | — |