Low Rank Approximations for Artificial Intelligence

In a context of data being collected and exploited at huge scales, designing efficient machine learning tools that capture the complexity of data is one of the most important challenges of the decade. Low-rank approximations are such tools, that look for information shared across all modes of a multiway array. Low-rank approximations are a principal tool in machine learning, however mostly in the realm of unsupervised learning. In particular, adding external information such as labeled data, a known dictionary of features or additional multimodal dataset raises challenging questions on how to rediscover low-rank approximations methods in the context of semi-supervised learning. Project LoRAiA will study the theoretical properties of such semi-supervised problems. LoRAiA will also produce efficient algorithms to solve the underlying existing and new optimization problems.