Kód: 52475772
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- ... celý popis
Angličtina
Nákupem získáte 77 bodů
Anotace knihy
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.
Parametry knihy
Zařazení knihy Knihy v angličtině Computing & information technology Computer science
774 Kč
AngličtinaOsobní odběr Praha, Brno a 47529 dalších
Copyright ©2008-26 nejlevnejsi-knihy.cz Všechna práva vyhrazenaSoukromíCookies
Vrácení do měsíce
571 999 099 (8-15.30h)Nákupní košík ( prázdný )