• Corpus ID: 126187322

Visualizing the decision-making process in deep neural decision forest

@inproceedings{Li2019VisualizingTD,
  title={Visualizing the decision-making process in deep neural decision forest},
  author={Shichao Li and K. Cheng},
  booktitle={CVPR Workshops},
  year={2019}
}
Deep neural decision forest (NDF) achieved remarkable performance on various vision tasks via combining decision tree and deep representation learning. In this work, we first trace the decision-making process of this model and visualize saliency maps to understand which portion of the input influence it more for both classification and regression problems. We then apply NDF on a multi-task coordinate regression problem and demonstrate the distribution of routing probabilities, which is vital… 

Figures and Tables from this paper

SegNBDT: Visual Decision Rules for Segmentation
TLDR
This work builds a hybrid neural-network and decision-tree model for segmentation that attains neural network segmentation accuracy and provides semi-automatically constructed visual decision rules such as "Is there a window?".
NBDT: Neural-Backed Decision Trees
TLDR
This work proposes Neural-Backed Decision Trees (NBDTs), modified hierarchical classifiers that use trees constructed in weight-space that achieve interpretability and neural network accuracy and match state-of-the-art neural networks on CIFar10, CIFAR100, TinyImageNet, and ImageNet.
Neural Prototype Trees for Interpretable Fine-grained Image Recognition
TLDR
The Neural Prototype Tree (ProtoTree), an intrinsically interpretable deep learning method for fine-grained image recognition that combines prototype learning with decision trees, and thus results in a globally interpretable model by design.
Exploring Intermediate Representation for Monocular Vehicle Pose Estimation
TLDR
A new learning-based framework to recover vehicle pose in SO(3) from a single RGB image that outperforms previous monocular RGB-based methods for joint vehicle detection and pose estimation on the KITTI benchmark, achieving performance even comparable to stereo methods.
Facial age estimation by deep residual decision making
TLDR
This paper incorporates residual learning into NDF and the resulting model achieves state-of-the-art level accuracy on three public age estimation benchmarks while requiring less memory and computation.

References

SHOWING 1-10 OF 11 REFERENCES
Deep Neural Decision Forests
TLDR
A novel approach that unifies classification trees with the representation learning functionality known from deep convolutional networks, by training them in an end-to-end manner by introducing a stochastic and differentiable decision tree model.
Interpreting CNNs via Decision Trees
TLDR
The proposed decision tree is a decision tree, which clarifies the specific reason for each prediction made by the CNN at the semantic level, and organizes all potential decision modes in a coarse-to-fine manner to explain CNN predictions at different fine-grained levels.
Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
TLDR
This paper addresses the visualisation of image classification models, learnt using deep Convolutional Networks (ConvNets), and establishes the connection between the gradient-based ConvNet visualisation methods and deconvolutional networks.
Deep Residual Learning for Image Recognition
TLDR
This work presents a residual learning framework to ease the training of networks that are substantially deeper than those used previously, and provides comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth.
Deep Regression Forests for Age Estimation
TLDR
The proposed Deep Regression Forests (DRFs), an end-to-end model for age estimation, connect the split nodes to a fully connected layer of a convolutional neural network (CNN) and deal with inhomogeneous data by jointly learning input-dependant data partitions at the split node and data abstractions at the leaf nodes.
Interpretable Convolutional Neural Networks
TLDR
A method to modify a traditional convolutional neural network into an interpretable CNN, in order to clarify knowledge representations in high conv-layers of the CNN, which can help people understand the logic inside a CNN.
The First 3 D Face Alignment in the Wild ( 3 DFAW ) Challenge
TLDR
The results suggest that 3D alignment from 2D video is feasible on a wide range of face orientations and suggest directions for further research.
The First 3D Face Alignment in the Wild (3DFAW) Challenge
TLDR
The results suggest that 3D alignment from 2D video is feasible on a wide range of face orientations and suggest directions for further research.
A Fast and Accurate Unconstrained Face Detector
TLDR
A new image feature called Normalized Pixel Difference (NPD) is proposed, computed as the difference to sum ratio between two pixel values, inspired by the Weber Fraction in experimental psychology, which is scale invariant, bounded, and able to reconstruct the original image.
Very Deep Convolutional Networks for Large-Scale Image Recognition
TLDR
This work investigates the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting using an architecture with very small convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers.
...
...