🏆 ArXiv Paper Topic Classification Challenge

Welcome to the ArXiv Paper Topic Classification Challenge!

Every day, thousands of scientific papers are uploaded to arXiv.org
, covering fields from computer science to mathematics and physics. With such a vast amount of research, automatically classifying papers into their correct categories is an essential step for organizing knowledge and making it more discoverable.

In this competition, you will build machine learning models that can predict the research category of an academic paper based only on its metadata.

🎯 Goal

Your task is to predict the category label (label) of each paper in the test set, given the provided information.
This is a multi-class text classification problem, using the top-N most common arXiv categories (e.g., cs.CL, math.CO, physics.optics, etc.).

Participants will be ranked based on their model’s performance on the hidden test set.