Adversarial Differentiable Rendering for Grasp Quality Convolutional Neural Networks published in ASME Journal on Autonomous Vehicle Systems (JAVS) October 2026. Work presents a novel differentiable implementation of the physics-based Canny-Ferrari grasp quality and a novel self-collision avoidance strategy to create a new adversarial attack based on 3D differentiable rendering to mislead a grasp quality neural network.