A Two-Stage Subspace Trust Region Approach for Deep Neural Network Training
This work addresses the challenge of efficient and robust neural network training for machine learning practitioners, though it appears incremental as it builds on existing second-order and trust region methods.
The paper tackles the problem of training feed-forward neural networks by introducing a novel second-order method that uses a low-dimensional subspace quadratic approximation and a two-stage trust region minimization, resulting in fast objective function decay and avoidance of saddle points without manual parameter tuning, demonstrated on benchmark datasets.
In this paper, we develop a novel second-order method for training feed-forward neural nets. At each iteration, we construct a quadratic approximation to the cost function in a low-dimensional subspace. We minimize this approximation inside a trust region through a two-stage procedure: first inside the embedded positive curvature subspace, followed by a gradient descent step. This approach leads to a fast objective function decay, prevents convergence to saddle points, and alleviates the need for manually tuning parameters. We show the good performance of the proposed algorithm on benchmark datasets.