DELPHI · Computing Answers to Complex Questions in Broad Domains

The explosion of information around us has democratized knowledge and transformed its availability forpeople around the world. Still, since information is mediated through automated systems, access is boundedby their ability to understand language.Consider an economist asking “What fraction of the top-5 growing countries last year raised their co2 emission?”.While the required information is available, answering such complex questions automatically isnot possible. Current question answering systems can answer simple questions in broad domains, or complexquestions in narrow domains. However, broad and complex questions are beyond the reach of state-of-the-art.This is because systems are unable to decompose questions into their parts, and find the relevant informationin multiple sources. Further, as answering such questions is hard for people, collecting large datasets to trainsuch models is prohibitive.In this proposal I ask: Can computers answer broad and complex questions that require reasoning overmultiple modalities? I argue that by synthesizing the advantages of symbolic and distributed representationsthe answer will be “yes”. My thesis is that symbolic representations are suitable for meaning composition, asthey provide interpretability, coverage, and modularity. Complementarily, distributed representations (learnedby neural nets) excel at capturing the fuzziness of language. I propose a framework where complex questionsare symbolically decomposed into sub-questions, each is answered with a neural network, and the final answeris computed from all gathered information.This research tackles foundational questions in language understanding. What is the right representationfor reasoning in language? Can models learn to perform complex actions in the face of paucity of data?Moreover, my research, if successful, will transform how we interact with machines, and define a role forthem as research assistants in science, education, and our daily life.