Our paper Enforcing the consensus between Trajectory Optimization and Policy Learning for precise robot control (authored by Quentin le Lidec) was accepted at ICRA 2023.