Papers

Fast-JEPA:低延迟潜在世界模型规划框架,推理速度达 INTACT 的 2.41 倍

Fast-JEPA: Advancing the Success-Latency Frontier for Latent World Models

Why we picked it

世界模型做规划通常要一步步递归推演,测试时很慢。这篇把候选动作序列的结果直接预测出来再比较,在四个环境上比 INTACT 快 2.41 倍,成功率还有 96.50%。做机器人控制和模型规划方向的开发者可以看看,代码已开源,附带一个新基准 DynaGoal。

Fast-JEPA 是一种潜在世界模型规划框架,在 PushT、Reacher、Cube、TwoRoom 四个环境上取得 96.50% 平均成功率,规划速度比 INTACT 快 2.41x。它的核心改进是重构预测计算方式:先构造目标导向的规划来引导搜索,然后直接预测完整候选动作序列的结果用于比较,避开传统方法在搜索细化时逐状态递归 rollout 的高开销。

Body

作者还发布了 DynaGoal 基准,用于评估世界模型规划在动态分布外条件、执行期限和受控响应延迟下的表现。Fast-JEPA 在 DynaGoal-Shift 上平均成功率为 73.83%,在 DynaGoal-Deadline 上为 87.67%。搜索预算扫描实验显示,其成功率和低延迟优势在不同候选数量与细化深度下均保持稳定。代码已在 GitHub 开源。

Original · arXiv cs.AI

Fast-JEPA: Advancing the Success-Latency Frontier for Latent World Models

Latent world models predict the outcomes of action sequences, but comparing candidate plans through recursive rollout remains expensive at test time. We introduce Fast-JEPA, a latent world model planning framework that advances the success-latency Pareto frontier of predictive search. Fast-JEPA reorganizes predictive computation to support effective action-sequence search at low planning cost. It constructs a goal-directed plan to guide search and directly predicts the outcomes of complete candidates for efficient comparison, avoiding recursive state rollout during refinement. Across PushT, Reacher, Cube, and TwoRoom, Fast-JEPA achieves 96.50% average success rate with 2.41x faster planning than INTACT. Search-budget sweeps show that its competitive success and low latency extend across candidate counts and refinement depths. We further introduce DynaGoal, a benchmark for evaluating world-model planning under dynamic out-of-distribution conditions, execution deadlines, and controlled response delays. Fast-JEPA achieves average success rates of 73.83% on DynaGoal-Shift and 87.67% on DynaGoal-Deadline, demonstrating the practical value of low-latency predictive planning under changing control conditions and execution-time constraints. Our code is available at https://github.com/Sofilyzia/Fast-JEPA.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Fast-JEPA, world model, INTACT, DynaGoal.

Source attribution: curated from arXiv cs.AI (@Yao Zhang, Bin Yu, Shijie Lian, Haishan Liu, Changti Wu, Yu Wu, Kai Chen).