Unverified50% confidenceFactExact time
This smart driving challenge was run on the Momenta R7 world model.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/16/2026
First Seen
Valid until: 1/14/2027
Sources
Related Entities
Related Claims
UnverifiedThe core idea of the Momenta R7 world model is to learn physical laws, common sense, and causality from massive real driving data, in order to improve environment understanding, prediction, and planning capabilities.72% similarUnverifiedApproval results from Human-in-the-Loop (HITL) can simultaneously serve as RLHF training data, forming a flywheel of continuously improving model capability.58% similarUnverified项目参考van Laatum等人的研究（arXiv:2607.07365），采用轨迹选择式RL-MPC方法：MPC生成若干候选控制序列，RL策略从中选择最优轨迹57% similarUnverifiedRLAIF通过让AI模型自动生成偏好标注数据，突破人工标注的速度瓶颈56% similarUnverifiedOllama v0.35.1-rc0 introduced an explicit model capabilities declaration mechanism to the model creation process via PR #18708.55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1046377API
curl https://kongchang.com/api/v1/knowledge/claims/1046377MCP
get_claim(id=1046377)