Search and exploration of scanned books

Abstract

In this demo, we present Proteus, a novel interface for interacting with multiple retrieval types extracted from scanned books provided by the Internet Archive. The primary purpose of Proteus is to provide a rich interactive experience for users to explore collections with automatically extracted and linked entity data. The system supports seamlessly shifting perspectives between books, entities, and topics. Proteus provides a starting point for a variety of exploratory search tasks.