Scientific Statement Classification over arXiv.org

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Ginev, Deyan, Miller, Bruce R.
Formato: Preprint
Publicado: 2019
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912283359182848
author Ginev, Deyan
Miller, Bruce R.
author_facet Ginev, Deyan
Miller, Bruce R.
contents We introduce a new classification task for scientific statements and release a large-scale dataset for supervised learning. Our resource is derived from a machine-readable representation of the arXiv.org collection of preprint articles. We explore fifty author-annotated categories and empirically motivate a task design of grouping 10.5 million annotated paragraphs into thirteen classes. We demonstrate that the task setup aligns with known success rates from the state of the art, peaking at a 0.91 F1-score via a BiLSTM encoder-decoder model. Additionally, we introduce a lexeme serialization for mathematical formulas, and observe that context-aware models could improve when also trained on the symbolic modality. Finally, we discuss the limitations of both data and task design, and outline potential directions towards increasingly complex models of scientific discourse, beyond isolated statements.
format Preprint
id arxiv_https___arxiv_org_abs_1908_10993
institution arXiv
publishDate 2019
record_format arxiv
spellingShingle Scientific Statement Classification over arXiv.org
Ginev, Deyan
Miller, Bruce R.
Computation and Language
Artificial Intelligence
Digital Libraries
We introduce a new classification task for scientific statements and release a large-scale dataset for supervised learning. Our resource is derived from a machine-readable representation of the arXiv.org collection of preprint articles. We explore fifty author-annotated categories and empirically motivate a task design of grouping 10.5 million annotated paragraphs into thirteen classes. We demonstrate that the task setup aligns with known success rates from the state of the art, peaking at a 0.91 F1-score via a BiLSTM encoder-decoder model. Additionally, we introduce a lexeme serialization for mathematical formulas, and observe that context-aware models could improve when also trained on the symbolic modality. Finally, we discuss the limitations of both data and task design, and outline potential directions towards increasingly complex models of scientific discourse, beyond isolated statements.
title Scientific Statement Classification over arXiv.org
topic Computation and Language
Artificial Intelligence
Digital Libraries
url https://arxiv.org/abs/1908.10993