Deep neural networks regularization for structured output prediction

Abstract

A deep neural network model is a powerful framework for learning representations. Usually, it is used to learn the relation x → y by exploiting the regularities in the input x. In structured output prediction problems, y is multi-dimensional and structural relations often exist between the dimensions. The motivation of this work is to learn the output dependencies that may lie in the output data in order to improve the prediction accuracy. Unfortunately, feedforward networks are unable to exploit the relations between the outputs. In order to overcome this issue, we propose in this paper a regularization scheme for training neural networks for these particular tasks using a multi-task framework. Our scheme aims at incorporating the learning of the output representation y in the training process in an unsupervised fashion while learning the supervised mapping function x → y. We evaluate our framework on a facial landmark detection problem which is a typical structured output task. We show over two public challenging datasets (LFPW and HELEN) that our regularization scheme improves the generalization of deep neural networks and accelerates their training. The use of unlabeled data and label-only data is also explored, showing an additional improvement of the results. We provide an opensource implementation 2 of our framework.

The Text Detection and Recognition Method for Electrical Nameplates Based on Deep Learning

“…The Seq2Seq model represented by encoding and decoding uses RNN or CNN to decode after the input image is encoded by convolution. Encoding is to convert an input sequence into a vector of constant length, and decoding is to decode an input vector of constant length into an output sequence [9] . The Seq2Seq model is commonly used in regression prediction.…”

Pet dog facial expression recognition based on convolutional neural network and improved whale optimization algorithm

“…Regularization techniques play a vital role in the development of machine learning models 41 , 42 . When fitting a model to some training dataset, the regularization is a common method to avoid over fitting.…”