Fork me on GitHub

Regression Concept Vectors (RCVs)

This repository contains the development of the interpretability and perturbation robustness for the analysis of histopathology images by convolutional neural networks.

2
mentions
1
contributor
56 commits | Last update: June 03, 2019

What Regression Concept Vectors (RCVs) can do for you

  • Extraction of concept measures from manual nuclei annotations
  • Correlation analysis between concepts and network prediction
  • Analysis of the concept learning dynamics during training
  • Perturbation analysis by bidirectional relevance scores

Recent work on Testing with Concept Activation Vectors (TCAV) proposed directional derivatives to quantify the in uence of user-defined concepts on the network output [1]. However, diagnostic concepts are often continuous measures that might be counter intuitive to describe by their presence or absence. In this paper, we extend TCAV from a classification problem to a regression problem by computing Regression Concept Vectors (RCVs). Instead of seeking a discriminator between two concepts, we seek the direction of greatest increase of the measures for a single continuous concept. We measure the relevance of a concept with bidirectional relevance scores, Br. The Br scores assume positive values when increasing values of the concept measures positively affect classification and negative in the opposite case.

Read more
Tags
No tags available
Programming Language
  • Python
License
  • MIT
Source code

Participating organizations

Contributors

  • Mara Graziani
    Haute École Spécialisée de Suisse Occidentale HES-SO Valais (HESSO)
Contact person
Mara Graziani
Haute École Spécialisée de Suisse Occidentale HES-SO Valais (HESSO)