Estimate, Don't Imitate：复用可微状态策略做视觉运动控制
Estimate, Don't Imitate: Reusing Differentiable State-Based Policies for Visuomotor Control
训练机械臂策略的新思路：不蒸馏专家，只学视觉补状态，真机 Panda 上 76% 成功率，做 sim-to-real 的朋友可以看看。
论文提出一种替代 teacher-student 蒸馏的方案：保留基于状态的专家策略，只训练视觉状态估计器来补全其缺失的状态输入。训练结合直接状态监督与通过冻结的可微专家反传的动作一致性损失，并用调度目标逐步强调影响专家动作的误差。在 5 个目标条件操作任务上，该方法持续优于对同一专家演示做像素到动作的直接模仿。在真实 Panda 机器人上实现 sim-to-real 迁移，成功率达 76%，且无需重训底层专家。
Estimate, Don't Imitate: Reusing Differentiable State-Based Policies for Visuomotor Control
Simulation-trained manipulation policies can exploit privileged state information to learn effective contact-rich behaviours, but deployment requires acting from partial observations such as noisy camera images. A common solution is teacher-student distillation, in which a visuomotor policy is trained to reproduce the actions of the privileged expert. This requires the student to jointly infer the task-relevant state and relearn the expert's action mapping that is already available. An alternative is to reuse the state-based expert and learn only a perceptual interface that reconstructs its missing state inputs. However, minimising the state estimate error alone does not necessarily minimise the downstream control error induced by these estimates. To bridge this gap, we train a visual state estimator using both direct state supervision and an action-consistency loss backpropagated through the frozen, differentiable expert. A scheduled objective first establishes a physically meaningful state estimate and progressively emphasises errors that affect the expert's actions. Across five goal-conditioned manipulation tasks, retaining the expert consistently outperforms direct pixel-to-action imitation from the same expert demonstration corpus. We further demonstrate sim-to-real transfer on a physical Panda robot, achieving 76% success without retraining the underlying expert.
