Oct 07, 2026 Q-Learning with Scalar Adjoint Matching: Cheaper Off-Policy RL for Flow Policies May 25, 2026 Trust Region Q Adjoint Matching: Stable Off-Policy RL for Flow Policies