Evoke combines an external camera-indexed geometry memory with a long-horizon teacher to produce a three-step world model that generates hour-scale video with bounded per-step cost.
Creativity Commonsense Controllability Human Fidelity Physics Model Composition Diversity Instance Preserv
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Alaya-EVOKE: From Linear-Scaling Supervision to Endless World
Evoke combines an external camera-indexed geometry memory with a long-horizon teacher to produce a three-step world model that generates hour-scale video with bounded per-step cost.