Programming by Teaching (Bloomsbury AI)

Speaker

Guillaume Bouchard

Abstract

Recent advances in machine learning and natural language processing open the door to new applications in many domains, including computer-assisted creativity. In this talk, we will review the various programming paradigms that have been used since the invention of computer. We show how it becomes possible for computers to learn and understand human language by reviewing the majors technical advances that are currently tackled by researcher: 1) grounding, i.e. the ability to refer to real objects in the world, 2) understanding explanations to learn with fewer data, and 3) generating explanations to better understand decisions made by the computer. Finally, a detailed overview of the recent advances in the statistical relational learning domain will be given.

SPEAKER BIO

Dr Bouchard is CEO and co-founder of Bloomsbury AI, an early-stage company automating expertise through natural language processing and machine learning approaches. He was previously senior researcher at UCL, focusing on statistical learning for large amounts of relational data. This includes text understanding, user modelling, and social media analytics. The main application areas of his work includes link prediction — predictive algorithms for relational data — for social media monitoring, transportation analytics, and content creation. Guillaume holds a PhD in statistics from INRIA Rhône-Alpes and Université Joseph Fourier (France), an engineering degree in mathematics from Institut National des Sciences Appliquées, and a masters in Applied Mathematics with a specialization in probability theory.