Unverified50% confidenceFactExact time
Training a neural network directly on consecutive, correlated experiences leads to catastrophic forgetting and training instability.
1
Sources
50%
Confidence
Long-term
Relevance
11/2/2026
First Seen
Sources
Related Entities
Related Claims
Unverified经验回放将交互经验存入固定大小缓冲区并随机采样小批量训练，打破连续经验的时间相关性，使神经网络训练更稳定78% similarUnverified解决灾难性遗忘的方案是在训练时混入10%-20%的原始预训练数据75% similarUnverifiedDirectly approximating optimal control laws with neural networks carries the risk of constraint violations, and traditional supervised learning requires pre-computing a large number of expensive NMPC solutions as training labels.74% similarUnverifiedWhen a neural network trained offline is deployed to a closed-loop control system, its actions change the system's state distribution, causing the data distribution at actual runtime to deviate from the offline data distribution at training time, that is, distribution shift.73% similarUnverifiedThe loss surface of multi-layer neural networks is filled with local minima and saddle points, and training requires optimization techniques such as momentum, adaptive learning rates, regularization, and batch normalization.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1125985API
curl https://kongchang.com/api/v1/knowledge/claims/1125985MCP
get_claim(id=1125985)