A Gradient Boosting Approach for Training Convolutional and Deep Neural Networks
This work addresses training efficiency and performance for neural networks in computer vision and tabular data, but it is incremental as it builds on existing gradient boosting and neural network methods.
The authors tackled the problem of training convolutional and deep neural networks by introducing gradient boosting-based procedures (GB-CNN and GB-DNN), which iteratively add dense layers to fit pseudo-residuals and freeze previous weights to prevent overfitting, resulting in superior classification accuracy compared to standard architectures on 2D-image and tabular datasets.
Deep learning has revolutionized the computer vision and image classification domains. In this context Convolutional Neural Networks (CNNs) based architectures are the most widely applied models. In this article, we introduced two procedures for training Convolutional Neural Networks (CNNs) and Deep Neural Network based on Gradient Boosting (GB), namely GB-CNN and GB-DNN. These models are trained to fit the gradient of the loss function or pseudo-residuals of previous models. At each iteration, the proposed method adds one dense layer to an exact copy of the previous deep NN model. The weights of the dense layers trained on previous iterations are frozen to prevent over-fitting, permitting the model to fit the new dense as well as to fine-tune the convolutional layers (for GB-CNN) while still utilizing the information already learned. Through extensive experimentation on different 2D-image classification and tabular datasets, the presented models show superior performance in terms of classification accuracy with respect to standard CNN and Deep-NN with the same architectures.