Soft Cosine Measure at ARQMath

This repository contains the math information retrieval (MIR) system for the ARQMath competition based on the tf-idf model and the soft cosine measure [[1][]]:

Joint word embeddings and soft cosine measure at ARQMath

The system build on the following other repositories:

• ARQMath-data-preprocessing: Scripts for producting preprocessed ARQMath competition datasets.
• ARQMath-eval: Python package for evaluating the performance of a MIR system on a number of tasks, including ARQMath.

Usage

To set up our system, execute the following commands:

git submodule update --init
pip install -r input_data/requirements.txt
pip install -r requirements.txt

Then, you can either:

1. download our results by executing the dvc pull command, or
2. reproduce our results by installing the dependencies of the ARQMath-data-preprocessing repository, reproducing or downloading the results in the ARQMath-data-preprocessing repository, and then executing the dvc repro command.

Citing

Text

NOVOTNÝ, Vít, Petr SOJKA, Michal ŠTEFÁNIK and Dávid LUPTÁK. Three is Better than One: Ensembling Math Information Retrieval Systems. CEUR Workshop Proceedings. Thessaloniki, Greece: M. Jeusfeld c/o Redaktion Sun SITE, Informatik V, RWTH Aachen., 2020, vol. 2020, No 2696, p. 1-30. ISSN 1613-0073.

BibTeX

@inproceedings{mir:mirmuARQMath2020,
  title = {{Three is Better than One}},
  author = {V\'{i}t Novotn\'{y} and Petr Sojka and Michal \v{S}tef\'{a}nik and D\'{a}vid Lupt\'{a}k},
  booktitle = {CEUR Workshop Proceedings: ARQMath task at CLEF conference},
  publisher = {CEUR-WS},
  address = {Thessaloniki, Greece},
  date = {22--25 September, 2020},
  year = 2020,
  volume = 2696,
  pages = {1--30},
  url = {http://ceur-ws.org/Vol-2696/paper_235.pdf},
}