Solving Venn-Diagram Based Comprehension Questions in Set Theory using Natural Language Processing

Abstract

Natural language problems have gained increased attention in the recent times. There are many special purposes, and carefully constructed evaluations driving the NLP research. Problem solving tests offer an interesting alternative to these evaluations. The primary goal of creating these problem solving tests is to evaluate the reading skills, and there by create bank of training materials and ranking procedures to match the existing measures of human performance. Solving Set Theory problems using NLP exposes once such research problem and helps creating an evaluation method for Natural Language Understanding Systems. This paper describes the possibility to challenge these systems to successively push higher performance levels up to an accuracy of 80% with high speed as an added advantage.