读论文-NN landscape

深度线性神经网络：每个local-min就是global-min

文献：

[6] P. Baldi and K. Hornik, “Neural networks and principal component analysis: 
Learning from examples without local minima,” Neural Netw., vol. 2, no. 1, pp. 
53–58, 1989. doi: 10.1016/0893-6080(89)90014-2.

[24] K. Kawaguchi, “Deep learning without poor local minima,” in Proc. Advances 
Neural Information Processing Systems, 2016, pp. 586–594.

[44] C. Yun, S. Sra, and A. Jadbabaie, “Global optimality conditions for deep neural 
networks,” in Proc. Int. Conf. Learning Representations, 2018.

global-min