Unverified50% confidenceTradeoffExact time
The cost of freezing the world model training strategy is: if the world model itself has bias, the policy will overfit to that bias, i.e., the phenomenon of 'model exploitation' occurs.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedWorld-model-based planning falls into two categories: Model Predictive Control (MPC) and training policies within simulation.73% similarUnverifiedDuring inference, model weights are frozen, and the model is merely managing context rather than converting present experience into persistent model knowledge.72% similarUnverified领先模型若被用于生成训练数据，可能将自身偏差系统性地传递给下一代，即'模型坍塌'风险71% similarUnverifiedDreamer's policy learning adopts the Actor-Critic framework: the Actor outputs an action distribution based on the imagined latent state s_t, and the Critic estimates the value of that state; gradients are backpropagated through the world model's latent trajectory, but the world model parameters are usually frozen to avoid the policy optimization signal contaminating the world model's learning objective, thereby maintaining training stability.71% similarUnverified模型漂移指模型上线后由于真实世界数据分布变化导致预测性能随时间下降的现象，应对手段包括持续监控、性能告警阈值和定期或触发式再训练。70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011066API
curl https://kongchang.com/api/v1/knowledge/claims/1011066MCP
get_claim(id=1011066)