Tuesday, September 12th, 2023
03:00pm - 04:00pm, in Wheatley 3-154-28

Neriman Tokcan

UMass Boston

Tensor Methods: Uncovering Insights from Multidimensional Data

Abstract: Tensors are multidimensional arrays that generalize matrices to higher dimensions. They can extract meaningful patterns, relationships, and dependencies from multi-modal data that may be difficult to discover using traditional matrix-based methods. As a result, tensor-based methods have become an increasingly important tool for the efficient representation and analysis of high-dimensional complex data. They have found applications in several domains, including machine learning, image processing, bioinformatics, physics, social computing, and chemistry. In this talk, we introduce tensor tools and discuss the latest developments in the field. Our focus will shift towards genomics applications, particularly emphasizing tumor microenvironment analysis, showcasing their capacity to unravel critical insights from complex biological data. While recognizing the impactful potential of tensor methods, we will also address their inherent limitations and challenges. Furthermore, we will touch upon potential future developments, offering a glimpse into the evolving landscape of tensor methods across multiple domains.