Contrastive Entity Coreference and Disambiguation for Historical Texts

By Abhishek Arora, Emily Silcock, Leander Heldring, Melissa Dell

Rating

1136
Battle Count: 103

Relevance

3/10
While not directly applicable to quantitative trading, the entity disambiguation techniques could potentially be adapted for processing financial news or reports

Implementation Complexity

7/10
Requires understanding of contrastive learning and bi-encoder architectures, as well as significant data processing

Reproducibility

4/5
Code and datasets are made publicly available, but some computational resources may be required for full reproduction

About this paper

Methodology: Contrastive Learning. Problem types: Natural Language Processing, Entity Disambiguation, Coreference Resolution.

The interactive Everscope explorer (charts, battles, favorites) loads below.