WorldCycle post-trains interactive video world models with reinforcement learning rewards for spatial closure and temporal consistency on reversible action cycles, reducing long-horizon drift and improving composite-action accuracy.
Nature , volume =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models
WorldCycle post-trains interactive video world models with reinforcement learning rewards for spatial closure and temporal consistency on reversible action cycles, reducing long-horizon drift and improving composite-action accuracy.