WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models

  • 类型：arxiv
  • 标识：2608.04964
  • 链接：https://arxiv.org/abs/2608.04964
  • 主分类：multimodal
  • 形态：method
  • 被引：0
  • 被引来源：Semantic Scholar
  • S2被引：0
  • 影响力被引：0
  • TLDR：WorldCycle is introduced, a self-verifiable RL framework that constructs closed action cycles and their repeated executions from ordinary action sequences, and optimizes two complementary rewards: a spatial closure reward enforcing symmetry between mirrored forward and reverse segments, and a temporal consistency reward aligning states across repeated cycle executions.
  • 待LLM分类：否
  • 标题中文：WorldCycle：面向长视野视频世界模型的自可验证强化学习
  • TLDR中文：提出 WorldCycle，一种自验证 RL 框架，从普通动作序列中构建闭合动作循环及其重复执行，并优化两个互补奖励：空间闭合奖励（强制镜像的前向与反向片段之间的对称性）以及时间一致性奖励（对齐多次循环执行间的状态）。
  • 来源文件：
  • /inbox/tom/_candidates/2026-08-06-agent-rag-longcontext-candidates.json
  • [S2 enrich]