Are All the Datasets in Benchmark Necessary? A Pilot Study of Dataset Evaluation for Text Classification
This work addresses the problem of inefficient benchmarking for researchers in NLP and ML by identifying redundant datasets, which could streamline evaluation processes.
The paper investigates whether all datasets in benchmarks are necessary by evaluating their ability to distinguish between top-performing systems, finding that some widely used datasets contribute little while less used ones show strong discriminative power, and it explores predicting dataset discrimination based on properties like average sentence length with promising preliminary results.
In this paper, we ask the research question of whether all the datasets in the benchmark are necessary. We approach this by first characterizing the distinguishability of datasets when comparing different systems. Experiments on 9 datasets and 36 systems show that several existing benchmark datasets contribute little to discriminating top-scoring systems, while those less used datasets exhibit impressive discriminative power. We further, taking the text classification task as a case study, investigate the possibility of predicting dataset discrimination based on its properties (e.g., average sentence length). Our preliminary experiments promisingly show that given a sufficient number of training experimental records, a meaningful predictor can be learned to estimate dataset discrimination over unseen datasets. We released all datasets with features explored in this work on DataLab: \url{https://datalab.nlpedia.ai}.