Unverified50% confidenceFactExact time
The real world is non-stationary, and a policy's performance during training may deteriorate sharply after the environment undergoes a distribution shift.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
Verified仅用成功轨迹训练的模型往往存在分布外泛化的严重缺陷，遇到工具返回错误时便丧失纠偏能力73% similarUnverifiedDistribution shift refers to the statistical distribution difference between the training environment and the deployment environment, and is the direct mechanism by which spurious correlations cause performance collapse.71% similarUnverifiedDuring cross-domain fine-tuning, an excessively large learning rate will destroy pre-trained weights, while an excessively small one will result in almost no learning, requiring a careful trade-off.71% similarUnverifiedThe core challenge of offline RL is distributional shift, which can lead to extrapolation errors that cause the learned policy to perform far worse than expected at deployment.70% similarUnverifiedThe understanding of text-to-image models is essentially large-scale statistical association, and generation quality drops significantly when dealing with long-tail concepts that appear rarely in the training set.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/911441API
curl https://kongchang.com/api/v1/knowledge/claims/911441MCP
get_claim(id=911441)