Skip to content
Texas A&M University

Events for 11/09/2020 from all calendars

Seminar in Random Tensors

iCal  iCal

Time: 11:00AM - 12:00PM

Location: zoom

Speaker: Liza Rebrova, UCLA

Title: About modewise tensor dimension reduction and fitting low-rank tensors

Abstract: It is not a secret that probabilistic view in general and random matrix theory in particular present amazing tools to understand large high-dimensional data. However, in many cases, one has to go beyond “simple” matrix models to correctly represent and treat the data. For example, inherently multimodal data is better represented with a tensor, that is, higher-order generalization of a matrix. Transition to more advanced data structures sometimes can survive re-using old algorithms, however, the development of the special tools that honor the full structure within the data pays off by making the algorithms both much more efficient and better interpretable. In this talk, I will focus on our new provable methods for modewise (structure preserving) tensor dimension reduction. I will also discuss its application to the tensor fitting problem and the connections to interpretable learning from multi-modal data through tensor decompositions.

Industrial and Applied Math

iCal  iCal

Time: 6:30PM - 7:30PM

Location: ZOOM

Speaker: Dr. Marylesa Howard, DOE Nevada National Security Site

Title: Emerging Scientific Research Opportunities in Nuclear Security

Abstract: The Department of Energy (DOE) employs scientists, mathematicians, engineers, and technicians to work on problems ranging from renewable energy resources to global climate change. However, unbeknownst to many people is the fact that the DOE is also the nation’s overseer of nuclear non-proliferation, nuclear emergency response, and nuclear power for the U.S. Navy. Furthermore, the DOE also oversees our nuclear weapons program, where the safety, security, and reliability of our nuclear stockpile are governed through scientific modeling, simulation, and experimentation, and decisions about the nuclear stockpile program are driven by data analysis. In light of this, it is imperative that analysis methods be mathematically justifiable, the data obey the assumptions of the method, and the information extracted from each diagnostic’s data be useful and relevant to the questions posed. Statistical methods are absolutely necessary in the analysis of data from each diagnostic to estimate quantities of interest and to help us understand corresponding uncertainties. In this presentation, some of the scientific research interests of the DOE will be highlighted, including work of Postdocs, Graduate students, Co-Op students, and summer interns, with a focus on measurement diagnostics and analysis for subcritical experiments in support of the Stockpile Stewardship Program at the Nevada National Security Site, the nation’s premier explosives laboratory.