Autorenfreundlich Bücher kaufen?!
Beschreibung
We address interpretable representation learning for motion forecasting in self-driving cars. Rather than treating transformers as black boxes, we develop methods to interpret and modify learned representations. We introduce self-supervised pre-training with interpretable objectives. Moreover, we probe latent spaces of forecasting models and reveal interpretable features, allowing us to make targeted interventions. Finally, we uncover retrocausal mechanisms, which enable goal-based instructions.
Details
| Verlag | KIT Scientific Publishing |
| Ersterscheinung | 13. Mai 2026 |
| Maße | 21 cm x 14.8 cm |
| Gewicht | 330 Gramm |
| Format | Softcover |
| ISBN-13 | 9783731514749 |
| Auflage | 1. Auflage |
| Seiten | 172 |
Schlagwörter
Bekannt durch
