Unwrapping The Black Box of Deep ReLU Networks: Interpretability, Diagnostics, and Simplification

TL;DR

This paper introduces a local linear representation approach to interpret, diagnose, and simplify deep ReLU networks, making them more transparent and understandable for critical applications.

Contribution

It develops a set of tools based on local linear models to interpret, diagnose, and simplify deep ReLU networks, with practical visualization and merging strategies.

Findings

Effective interpretability and diagnostics demonstrated on benchmark datasets.

Successful network simplification via merging strategy.

Case study in credit risk assessment validates practical utility.

Abstract

The deep neural networks (DNNs) have achieved great success in learning complex patterns with strong predictive power, but they are often thought of as "black box" models without a sufficient level of transparency and interpretability. It is important to demystify the DNNs with rigorous mathematics and practical tools, especially when they are used for mission-critical applications. This paper aims to unwrap the black box of deep ReLU networks through local linear representation, which utilizes the activation pattern and disentangles the complex network into an equivalent set of local linear models (LLMs). We develop a convenient LLM-based toolkit for interpretability, diagnostics, and simplification of a pre-trained deep ReLU network. We propose the local linear profile plot and other visualization methods for interpretation and diagnostics, and an effective merging strategy for network simplification. The proposed methods are demonstrated by simulation examples, benchmark datasets, and a real case study in home lending credit risk assessment.