Don't Fear the Bit Flips: Optimized Coding Strategies for Binary Classification

TL;DR

This paper introduces new coding strategies to protect binary classifiers from noise-induced feature corruption, focusing on maximizing classifier output reliability rather than traditional data preservation methods.

Contribution

It proposes the same classification probability (SCP) as a measure of classifier robustness and develops low-complexity estimates and coding schemes tailored to enhance SCP.

Findings

SCP effectively quantifies classifier output distortion.

Error-correcting codes can be optimized for classifier robustness.

Proposed methods improve classifier reliability under noise.

Abstract

After being trained, classifiers must often operate on data that has been corrupted by noise. In this paper, we consider the impact of such noise on the features of binary classifiers. Inspired by tools for classifier robustness, we introduce the same classification probability (SCP) to measure the resulting distortion on the classifier outputs. We introduce a low-complexity estimate of the SCP based on quantization and polynomial multiplication. We also study channel coding techniques based on replication error-correcting codes. In contrast to the traditional channel coding approach, where error-correction is meant to preserve the data and is agnostic to the application, our schemes specifically aim to maximize the SCP (equivalently minimizing the distortion of the classifier output) for the same redundancy overhead.