Simulations and experimental processes in scientific and engineering applications often generate large amount of data. Data can be leveraged in a variety of Scientific Machine Learning (SciML) workflows, but data/dimension reduction strategies need to be first employed to reduce/compress, curate, and process these data. In this minisymposium, we welcome contributions focused on state-of-the-art strategies for dimension reduction and data compression with a particular emphasis on approaches relevant to SciML. In particular, we invite contributions that address one or more of the following topics: i) data compression and dimension reduction strategies that account for the preservation of relevant physical features; ii) approaches for the identification of low-dimensional structures and manifolds that can scale to large and/or distributed datasets; iii) data-driven approaches to data and dimension reduction that can leverage information sources characterized by varying degrees of accuracy and fidelity. We welcome contributions focused on algorithmic and/or theoretical advancements, but advanced deployments of these techniques to large scale scientific and engineering applications are also in scope. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525