LASSO: A Tool for Surfing the Answer Net

Abstract

This paper presents the architecture, operation and results obtained with the Lasso system developed in the Natural Language Processing Laboratory at SMU. The system relies on a combination of syntactic and semantic techniques, and lightweight abductive inference to find answers. The search for the answer is based on a novel form of indexing called paragraph indexing. A score of 55.5% for short answers and 64.5% for long answers was achieved. Background Finding the answer to a question by returning a small fragment of a text, where the answer actually lies, is profoundly di erent from the task of information retrieval (IR) or information extraction (IE). Current IR systems allow us to locate full documents that might contain the pertinent information, leaving it to the user to extract the answer from a ranked list of texts. In contrast, IE systems extract the information of interest, provided it has been presented in a prede ned, target representation, known as template. The i...