The loss surface of deep linear networks viewed through the algebraic geometry lens

no code implementations • 17 Oct 2018 • Dhagash Mehta, Tianran Chen, Tingting Tang, Jonathan D. Hauenstein

By using the viewpoint of modern computational algebraic geometry, we explore properties of the optimization landscapes of the deep linear neural network models.

Fixed Points of Belief Propagation -- An Analysis via Polynomial Homotopy Continuation

no code implementations • 20 May 2016 • Christian Knoll, Franz Pernkopf, Dhagash Mehta, Tianran Chen

Moreover, we show that this fixed point gives a good approximation, and the NPHC method is able to obtain this fixed point.