Unverified50% confidenceSolutionExact time
The first attempt at the reproduction experiment failed to observe the degradation phenomenon because too few training epochs were used (the model was still in an underfitting state), indicating that sufficient convergence during training is a prerequisite for observing the degradation phenomenon.
1
Sources
50%
Confidence
Long-term
Relevance
8/15/2026
First Seen
Sources
深度网络退化问题实验复现：残差连接为何能拯救深层网络
Related Claims
Unverified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69% similarVerifiedOverfitting refers to the phenomenon where a model performs excellently on training data but its performance drops significantly on unseen test data.68% similarUnverifiedThe reproduction effort employs a scaled validation method, verifying the trends and relative relationships claimed in papers under smaller datasets, fewer training epochs, or lower model parameter counts.67% similarUnverifiedBehavior Cloning (BC) suffers from the covariate shift problem: during training the model only sees the state distribution of expert trajectories, and at deployment small errors can drive the robot into states not covered by the training set, leading to cascading failures.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/750228API
curl https://kongchang.com/api/v1/knowledge/claims/750228MCP
get_claim(id=750228)