Unverified50% confidenceFactExact time
行为克隆（BC）存在复合误差与分布偏移问题，智能体偏离训练分布后预测误差会指数级累积
1
Sources
50%
Confidence
Long-term
Relevance
7/19/2026
First Seen
Sources
Related Claims
VerifiedBehavioral cloning suffers from the distribution shift problem, where the model tends to make errors and accumulate mistakes on states outside the training distribution, a phenomenon referred to as compounding error in the robotics control field.82% 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.82% similarUnverifiedAn intermittently faulty memory module may alternately produce correct and incorrect computation results, causing sporadic deviations in gradient calculations during neural network training, affecting model convergence without triggering hardware alerts.82% similarUnverifiedBehavioral cloning simplifies the problem to supervised learning, but suffers from distribution shift, where differences between the state distributions at training time and execution time lead to error accumulation.81% similarUnverifiedAgents face the problem of error cascading: judgment errors in early steps are amplified exponentially with the number of iterations.78% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/558812API
curl https://kongchang.com/api/v1/knowledge/claims/558812MCP
get_claim(id=558812)