Deep&Wide Learning

论文解释：https://blog.csdn.net/u010352603/article/details/80590129#11-memorization-和-generalization
                   https://xueqiu.com/9217191040/110449554

训练方法：小批量随机梯度下降https://blog.csdn.net/xiang_freedom/article/details/78395145

AdaGrad：https://blog.csdn.net/program_developer/article/details/80756008