An ALBERT-based Similarity Measure for Mathematical Answer Retrieval
Publikation: Beitrag in Buch/Konferenzbericht/Sammelband/Gutachten › Beitrag in Konferenzband › Beigetragen › Begutachtung
Beitragende
Abstract
Mathematical Language Processing (MLP) deals with the automated processing and analysis of mathematical documents and relies heavily on good representations of mathematical symbols and texts. The aim of this work is to explore the modeling capabilities of state-of-the-art unsupervised deep learning methods to create such representations. Therefore, we pre-trained different instances of an ALBERT model on Mathematics StackExchange data and fine-tuned it on the task of Mathematical Answer Retrieval. Our evaluation shows that ALBERT outperforms all previous systems and is on par with current state-of-the-art systems for math retrieval indicating strong capabilities of modeling mathematical posts. This implies that our approach can also be beneficial to various other tasks in MLP such as automatic proof checking or summarization of scientific texts.
Details
Originalsprache | Englisch |
---|---|
Titel | SIGIR '21: Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval |
Herausgeber (Verlag) | Association for Computing Machinery, Inc |
Seiten | 1593-1597 |
Seitenumfang | 5 |
ISBN (elektronisch) | 978-1-4503-8037-9 |
Publikationsstatus | Veröffentlicht - 11 Juli 2021 |
Peer-Review-Status | Ja |
Publikationsreihe
Reihe | IR: Research and Development in Information Retrieval |
---|
Konferenz
Titel | 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2021 |
---|---|
Dauer | 11 - 15 Juli 2021 |
Stadt | Virtual, Online |
Land | Kanada |
Externe IDs
Scopus | 85111688215 |
---|---|
ORCID | /0000-0001-8107-2775/work/142253439 |
Schlagworte
ASJC Scopus Sachgebiete
Schlagwörter
- information retrieval, mathematical language processing