Watson Jeopardy! Architecture

Why It Was Important

IBM began developing Watson to process unstructured text data and answer complex questions fast enough to compete on the gameshow Jeopardy! Utilizing the DeepQA architecture—firing hundreds of simultaneous hypothesis-generating algorithms and scoring their confidence—it proved the viability of massive, pipeline-based NLP over massive text corpora.

Who Invented It

IBM Research

A large team led by David Ferrucci.

Applications

• Question Answering
• Healthcare Data Analysis
• Enterprise AI