Verified70% confidenceFactExact time
Behavioral 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.
4
Sources
70%
Confidence
Long-term
Relevance
7/5/2026
First Seen
Sources
Related Claims
VerifiedBehavioral 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.85% 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.83% similarUnverified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82% similarVerified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79% similarUnverifiedBehavioral Cloning is susceptible to the compounding error problem, whereby minor deviations during policy execution accumulate continuously, causing the trajectory to deviate from the demonstration distribution.77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/112220API
curl https://kongchang.com/api/v1/knowledge/claims/112220MCP
get_claim(id=112220)