English

SetConv: A New Approach for Learning from Imbalanced Data

Information Retrieval 2021-04-14 v1 Artificial Intelligence Machine Learning

Abstract

For many real-world classification problems, e.g., sentiment classification, most existing machine learning methods are biased towards the majority class when the Imbalance Ratio (IR) is high. To address this problem, we propose a set convolution (SetConv) operation and an episodic training strategy to extract a single representative for each class, so that classifiers can later be trained on a balanced class distribution. We prove that our proposed algorithm is permutation-invariant despite the order of inputs, and experiments on multiple large-scale benchmark text datasets show the superiority of our proposed framework when compared to other SOTA methods.

Keywords

Cite

@article{arxiv.2104.06313,
  title  = {SetConv: A New Approach for Learning from Imbalanced Data},
  author = {Yang Gao and Yi-Fan Li and Yu Lin and Charu Aggarwal and Latifur Khan},
  journal= {arXiv preprint arXiv:2104.06313},
  year   = {2021}
}

Comments

Accepted by EMNLP 2020 (11 pages, 9 figures)

R2 v1 2026-06-24T01:07:48.324Z