TY - RPRT TI - Right for the Right Reasons: Training Differentiable Models by Constraining their Explanations AU - Andrew Slavin Ross AU - Michael C. Hughes AU - Finale Doshi-Velez PY - 2017 UR - https://arxiv.org/abs/1703.03717 ID - 1703.03717 ER -