Deep Learning without Poor Local Minima — mentioned by visarga

Tue, 31 May 2016 04:51:52 GMT

What I knew was that local minima are not that problematic when the state space is highly dimensional. Only saddle points appear in high dimensional spaces. The 2D example is not realistic. Here is a quote from a paper I randomly sampled on arxiv: > For an expected loss function of a deep nonlinear neural network, we prove the following statements under the independence assumption adopted from recent work: 1) the function is non-convex and non-concave, 2) every local minimum is a global minimum, 3) every critical point that is not a global minimum is a saddle point, and 4) the property of…

Paper and discussion history. A mention is not an endorsement.

Deep Learning without Poor Local Minima — HN submission

Tue, 24 May 2016 20:14:00 GMT

Deep Learning Without Poor Local Minima

123 points; 32 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Deep Learning without Poor Local Minima · HackerPapers discussions

Deep Learning without Poor Local Minima — mentioned by visarga

Deep Learning without Poor Local Minima — HN submission