# Fine-Grained Temporal Relation Extraction

## TL;DR

This paper introduces a new semantic framework for modeling detailed temporal relations and durations between events, supported by a large dataset and effective prediction models, advancing temporal understanding in NLP.

## Contribution

It proposes a novel framework for fine-grained temporal relation modeling, creates the largest related dataset, and demonstrates effective joint prediction methods.

## Key findings

- Strong predictive results on the dataset
- Effective transfer-learning approach for categorical relations
- Largest dataset covering the entire UD English Web Treebank

## Abstract

We present a novel semantic framework for modeling temporal relations and event durations that maps pairs of events to real-valued scales. We use this framework to construct the largest temporal relations dataset to date, covering the entirety of the Universal Dependencies English Web Treebank. We use this dataset to train models for jointly predicting fine-grained temporal relations and event durations. We report strong results on our data and show the efficacy of a transfer-learning approach for predicting categorical relations.