Building Watson: An Overview of the DeepQA Project

Abstract

IBM Research undertook a challenge to build a computer system that could compete at the human champion level in real time on the American TV Quiz show, Jeopardy! The extent of the challenge includes fielding a real-time automatic contestant on the show, not merely a laboratory exercise. The Jeopardy! Challenge helped us address requirements that led to the design of the DeepQA architecture and the implementation of Watson. After 3 years of intense research and development by a core team of about 20 researches, Watson is performing at human expert-levels in terms of precision, confidence and speed at the Jeopardy! Quiz show. Our results strongly suggest that DeepQA is an effective and extensible architecture that may be used as a foundation for combining, deploying, evaluating and advancing a wide range of algorithmic techniques to rapidly advance the field of QA.

Cognitive computing for coalition situational understanding

“…Natural language processing (NLP) A key element of IBM's original Watson [2], NLP is an essential component of CC systems that must process textual data and/or interact with human users in natural language.…”

“…While there is no generally agreed definition of the term, CC is characterised by a systems-oriented view in which multiple AI and/or SP technologies are integrated to address problems generally involving sizeable volumes of data. The archetypal CC system was IBM's original Watson, that integrated AI techniques including natural language processing, human-computer interaction, and knowledge representation and reasoning to achieve high performance in open-domain question answering [2]. Since the original Watson project, the concept of CC has broadened somewhat, such that we can identify a wider range of AI and SP capabilities occurring in cognitive systems.…”

Reconstructing Provenance

“…Inspired by the DeepQA approach of IBM Watson [11], we aim at developing a multi-signal pipeline, which will combine several signals representing evidence on the relationships between files and propose a ranked list of plausible reconstructions. Our multi-signal pipeline consists of four stages, each containing several components that can be executed in parallel:…”

A Natural Language Interface for Data Warehouse Question Answering

“…[10] is a system that can be used with any database, however it requires resources in addition to the database. Recently, [6] proposed a Q&A system dedicated to enterprise questions but also to general domain questions.…”