Preventing RNN from Using Sequence Length as a Feature

TL;DR

This paper identifies a critical flaw in RNNs where they exploit sequence length as a classification feature, leading to brittle models, and proposes a regularization-based solution to mitigate this issue.

Contribution

It reveals the problem of RNNs using sequence length as a shortcut and introduces a simple weight decay regularization method to prevent this behavior.

Findings

RNNs can rely on sequence length for classification

Regularization reduces length-based bias in RNNs

Models become more robust to concept drift

Abstract

Recurrent neural networks are deep learning topologies that can be trained to classify long documents. However, in our recent work, we found a critical problem with these cells: they can use the length differences between texts of different classes as a prominent classification feature. This has the effect of producing models that are brittle and fragile to concept drift, can provide misleading performances and are trivially explainable regardless of text content. This paper illustrates the problem using synthetic and real-world data and provides a simple solution using weight decay regularization.