Extending the Step-Size Restriction for Gradient Descent to Avoid Strict Saddle Points.

Hayden Schaeffer Scott G. McCalla

Published in: SIAM J. Math. Data Sci. (2020)

Keyphrases

step size
saddle points
cost function
stochastic gradient descent
scale space
convergence rate
convergence speed
critical points
objective function
saddle point
loss function
neural network
pattern recognition
image analysis
linear programming
genetic algorithm ga