Query-Focused Scenario Construction

TL;DR

This paper presents a query-focused system that constructs compatible event scenarios from conflicting news accounts using one-class clustering, evaluated on synthetic and real-world data, outperforming previous methods.

Contribution

It introduces a neural network-based approach for incremental scenario construction that accounts for event compatibility and order, with scalable synthetic training.

Findings

Neural network models significantly improve scenario construction accuracy.

Synthetic data training enhances model scalability and performance.

The system outperforms baseline methods on real-world news scenarios.

Abstract

The news coverage of events often contains not one but multiple incompatible accounts of what happened. We develop a query-based system that extracts compatible sets of events (scenarios) from such data, formulated as one-class clustering. Our system incrementally evaluates each event's compatibility with already selected events, taking order into account. We use synthetic data consisting of article mixtures for scalable training and evaluate our model on a new human-curated dataset of scenarios about real-world news topics. Stronger neural network models and harder synthetic training settings are both important to achieve high performance, and our final scenario construction system substantially outperforms baselines based on prior work.