Zhifeng Kong, Convergence Analysis of Training Two-Hidden-Layer Partially Over-Parameterized ReLU Networks via Gradient Descent. journal = {International Journal of Computer and Information Engineering}, [online]. World Academy of Science, Engineering and Technology. May 2020, vol. 162(6). 166 - 177 [viewed 28 April 2024]. Available from: https://publications.waset.org/pdf/10011232.