Detection Accuracy for Evaluating Compositional Explanations of Units

09/16/2021
by   Sayo M. Makinwa, et al.
0

The recent success of deep learning models in solving complex problems and in different domains has increased interest in understanding what they learn. Therefore, different approaches have been employed to explain these models, one of which uses human-understandable concepts as explanations. Two examples of methods that use this approach are Network Dissection and Compositional explanations. The former explains units using atomic concepts, while the latter makes explanations more expressive, replacing atomic concepts with logical forms. While intuitively, logical forms are more informative than atomic concepts, it is not clear how to quantify this improvement, and their evaluation is often based on the same metric that is optimized during the search-process and on the usage of hyper-parameters to be tuned. In this paper, we propose to use as evaluation metric the Detection Accuracy, which measures units' consistency of detection of their assigned explanations. We show that this metric (1) evaluates explanations of different lengths effectively, (2) can be used as a stopping criterion for the compositional explanation search, eliminating the explanation length hyper-parameter, and (3) exposes new specialized units whose length 1 explanations are the perceptual abstractions of their longer explanations.

READ FULL TEXT

page 2

page 7

page 8

research
06/24/2020

Compositional Explanations of Neurons

We describe a procedure for explaining neurons in deep representations b...
research
10/09/2021

Self-explaining Neural Network with Plausible Explanations

Explaining the predictions of complex deep learning models, often referr...
research
05/27/2018

Semantic Explanations of Predictions

The main objective of explanations is to transmit knowledge to humans. T...
research
09/15/2017

Embedding Deep Networks into Visual Explanations

In this paper, we propose a novel explanation module to explain the pred...
research
04/21/2019

GAN-based Generation and Automatic Selection of Explanations for Neural Networks

One way to interpret trained deep neural networks (DNNs) is by inspectin...
research
09/01/2020

Learning explanations that are hard to vary

In this paper, we investigate the principle that `good explanations are ...
research
11/04/2019

Learning to Annotate: Modularizing Data Augmentation for TextClassifiers with Natural Language Explanations

Deep neural networks usually require massive labeled data, which restric...

Please sign up or login with your details

Forgot password? Click here to reset