Machine Learning Data Analysis Project Presentation I – Rescheduled

Standardized tests have often been proposed as “drivers for progress in AI” (Clark et. al. 2016). These include reading comprehension tests (Richardson et. al. 2013), science question answering (Clark et. al. 2015), IQ tests, entrance exams, etc. These tests are widely accessible, easily comprehensible, clearly measurable, and offer a graduated progression from simple tasks to those requiring deep understanding of the world. This makes them perfect testbeds for research in knowledge-driven AI. In this talk, I will describe our approach to solve two of these standardized tests  – namely, reading comprehension and elementary science question answering. Both these tests require a significant level of language understanding, background knowledge modeling and reasoning. We posit that there is a hidden (latent) structure that explains the relation between the question, the correct answer, and the…


Link to Full Article: Machine Learning Data Analysis Project Presentation I – Rescheduled