Unverified70% confidenceFactExact time
Behavior 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.
2
Sources
70%
Confidence
Long-term
Relevance
7/14/2026
First Seen
Sources
Vizuara机器人学习课程笔记获取指南与免费资源推荐
redditr/deeplearning7/12/2026
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.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81% similarUnverifiedIn a closed-loop trajectory, a single action deviation puts the robot into a state outside the training distribution, and each subsequent step compounds error on top of the mistake, ultimately causing the entire episode to collapse—a failure that may show up as only a few percentage points of decline in average success rate.78% similarVerifiedDrift detection involves a trade-off: setting the threshold too sensitively leads to frequent false alarms and unnecessary retraining, while setting it too loosely may allow model performance to continuously degrade without being detected.77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/503251API
curl https://kongchang.com/api/v1/knowledge/claims/503251MCP
get_claim(id=503251)