CNN-RNN: a large-scale hierarchical image classification framework

Relation Classification: CNN or RNN?

Convolutional neural networks (CNN) have delivered competitive performance on relation classification, without tedious feature engineering. A particular shortcoming of CNN, however, is that it is less powerful in modeling longspan relations. This paper presents a model based on recurrent neural networks (RNN) and compares the capabilities of CNN and RNN on the relation classification task. We c...

Regularization Framework for Large Scale Hierarchical Classification

In this paper, we propose a hierarchical regularization framework for large-scale hierarchical classification. In our framework, we use the regularization structure to share information across the hierarchy and enforce similarity between class-parameters that are located nearby in the hierarchy. To address the computational issues that arise, we propose a parallel-iterative optimization scheme ...

Hierarchical Learning for Large-Scale Image Classification via CNN and Maximum Confidence Path

Large Scale Image Classification

We consider Multinomial Logistic Regression for large scale image classification. The model is trained using 1,000,000 images from Image Net Large Scale Visual Recognition Challenge 2010[2] . We train five models over different subset of sampled training observations. Finally we combine all the five models to obtain the test data classification. The combined classifier gives good performance. W...

HD-CNN: Hierarchical Deep Convolutional Neural Network for Image Classification

Improve classification accuracy of deep CNNs using hierarchical classification scheme.  Group classes based on confusion matrix.  Use networks of identical topology at various levels.