PerspectroScope: A Window to the World of Diverse Perspectives

TL;DR

PerspectroScope is a web-based system that enables users to explore diverse perspectives on natural language claims by extracting, visualizing, and supporting various viewpoints with evidence, leveraging recent NLP advances.

Contribution

It introduces a novel platform combining retrieval and textual entailment classifiers to visualize multiple perspectives on claims, with adaptive learning from user feedback.

Findings

Effective visualization of multiple perspectives

Improved coverage through adaptive mechanisms

Accessible via GitHub for community use

Abstract

This work presents PerspectroScope, a web-based system which lets users query a discussion-worthy natural language claim, and extract and visualize various perspectives in support or against the claim, along with evidence supporting each perspective. The system thus lets users explore various perspectives that could touch upon aspects of the issue at hand.The system is built as a combination of retrieval engines and learned textual-entailment-like classifiers built using a few recent developments in natural language understanding. To make the system more adaptive, expand its coverage, and improve its decisions over time, our platform employs various mechanisms to get corrections from the users. PerspectroScope is available at github.com/CogComp/perspectroscope.