# Deep Hierarchical Machine: a Flexible Divide-and-Conquer Architecture

## TL;DR

The paper introduces Deep Hierarchical Machine (DHM), a flexible divide-and-conquer deep model with probabilistic routing and pruning, optimized for classification and regression tasks with sparse feature extraction.

## Contribution

It presents a novel DHM architecture that combines stochastic routing with probabilistic pruning and sparse convolution, enhancing efficiency and flexibility over existing models.

## Key findings

- DHM outperforms previous architectures on image classification tasks.
- DHM demonstrates effective face alignment results.
- Sparse convolution improves computational efficiency.

## Abstract

We propose Deep Hierarchical Machine (DHM), a model inspired from the divide-and-conquer strategy while emphasizing representation learning ability and flexibility. A stochastic routing framework as used by recent deep neural decision/regression forests is incorporated, but we remove the need to evaluate unnecessary computation paths by utilizing a different topology and introducing a probabilistic pruning technique. We also show a specified version of DHM (DSHM) for efficiency, which inherits the sparse feature extraction process as in traditional decision tree with pixel-difference feature. To achieve sparse feature extraction, we propose to utilize sparse convolution operation in DSHM and show one possibility of introducing sparse convolution kernels by using local binary convolution layer. DHM can be applied to both classification and regression problems, and we validate it on standard image classification and face alignment tasks to show its advantages over past architectures.