@misc{17727,
  author = {Evrim Ataman},
  title = {Coupled Matrix and Tensor Factorizations {\textendash} Improving our Understanding of Complex Systems Through the Analysis of Temporal and Multimodal Data},
  abstract = {There is an emerging need to jointly analyze multimodal data sets and capture the underlying patterns in order to extract insights about complex systems. For instance, joint analysis of omics data (e.g., metabolomics, microbiome, genomics) holds the promise to bring together genomics, phenomics and environment in the quest for precision health. Such data sets are heterogeneous {\textendash} they are a collection of static and dynamic data. Dynamic data can often be arranged as a higher-order tensor (e.g., subjects by metabolites by time) while static data can be a matrix (e.g., subjects by genes). Tensor factorizations have been successfully used to reveal the underlying patterns in higher-order data, and extended to joint analysis of multimodal data through coupled matrix and tensor factorizations (CMTF). However, integrating heterogeneous data sets has still many challenges, especially when the goal is to capture interpretable (time-evolving) patterns. In this talk, we discuss CMTF models and algorithms for temporal and multimodal data mining. We focus on a flexible, accurate and computationally efficient framework (based on Alternating Optimization and Alternating Direction Method of Multipliers) that facilitates the use of a variety of constraints, loss functions and couplings with linear transformations when fitting CMTF models. Through various applications, we discuss the advantages and limitations of available CMTF methods.},
  year = {2024},
  journal = {SIAM Conference on Applied Linear Algebra, Paris, France},
}