TY - RPRT TI - Analysis of the expected $L_2$ error of an over-parametrized deep neural network estimate learned by gradient descent without regularization AU - Selina Drews AU - Michael Kohler PY - 2023 UR - https://arxiv.org/abs/2311.14609 ID - 2311.14609 ER -