On Evaluation of Document Classification using RVL-CDIP

TL;DR

This paper critically examines the RVL-CDIP benchmark for document classification, revealing significant issues like label noise and data overlap, and advocates for a new, more reliable benchmark to advance the field.

Contribution

The paper identifies key flaws in RVL-CDIP and provides recommendations for developing a more accurate and diverse document classification benchmark.

Findings

Label noise estimated at 8.1% in RVL-CDIP

Test and train data overlap inflate performance metrics

Presence of personally identifiable information in the dataset

Abstract

The RVL-CDIP benchmark is widely used for measuring performance on the task of document classification. Despite its widespread use, we reveal several undesirable characteristics of the RVL-CDIP benchmark. These include (1) substantial amounts of label noise, which we estimate to be 8.1% (ranging between 1.6% to 16.9% per document category); (2) presence of many ambiguous or multi-label documents; (3) a large overlap between test and train splits, which can inflate model performance metrics; and (4) presence of sensitive personally-identifiable information like US Social Security numbers (SSNs). We argue that there is a risk in using RVL-CDIP for benchmarking document classifiers, as its limited scope, presence of errors (state-of-the-art models now achieve accuracy error rates that are within our estimated label error rate), and lack of diversity make it less than ideal for benchmarking. We further advocate for the creation of a new document classification benchmark, and provide recommendations for what characteristics such a resource should include.