Unverified50% confidenceFactExact time
目标网络维护一份参数更新滞后的网络副本来计算TD目标值，避免训练发散
1
Sources
50%
Confidence
Long-term
Relevance
8/26/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedDQN introduced two key innovations—experience replay and target networks—where the former breaks the strong temporal correlation in sequential data to stabilize training, and the latter prevents training oscillation and divergence.68% similarUnverified强化学习中策略网络通过最大化累积奖励来优化参数，具有信息瓶颈特性，倾向于只编码对任务目标有用的信息65% similarUnverifiedA checkpoint refers to a snapshot of model parameters saved at a specific step during large model training, used to prevent loss of progress due to training interruptions62% similarUnverifiedTrigger-based retraining (based on performance degradation monitoring and data drift detection) can replace scheduled full retraining, directing computational resources precisely where they are truly needed62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/805335API
curl https://kongchang.com/api/v1/knowledge/claims/805335MCP
get_claim(id=805335)