I am Luella Fu, an assistant professor in the Mathematics Department, at San Francisco State University. I have been here since 2018 and have benefited immensely from supportive colleagues and diligent students. I deeply appreciate this community.
My methodological focuses are in nonparametric empirical Bayes methods on multiple testing and other inference using heteroscedastic data . Like much of statistics, these methodologies grew from difficulties in real data. Specifically, genome sequencing technology allowed scientists to contrast thousands of genes ("large-scale" data) between sick and healthy patients for a wide variety of illnesses, but the scientists needed a way to accurately detect which genes were related to illness and which were not ("multiple testing"). As methods and results multiplied, it became increasingly important to obtain replicable results, and part of solving this lies in accounting for the variability of data ("heteroscedasticity"). My enthusiasm for this field comes from its potential to expand into the technological realm, where multiple testing frameworks can provide mathematically-grounded algorithms for fraud detection and quality control. There is still a lot of work to be done here both algorithmically and theoretically because the scale of online data is such that the error rates need to be much lower than what is traditionally acceptable. There is also a lot of opportunity in online data because there is so much information that can be used to create models that require fewer statistical assumptions but that still have mathematical properties ("nonparametric empirical Bayes methods"). This is an amazing field with so many applications to large-scale data and so much still to develop in terms of theoretically-backed statistical techniques.I am also interested in how statistics are used in other applications, and have worked on medical and social network problems. Click for my CV.
Teaching is also important to me. First, for any students reading this who are trying to decide on classes, the first big idea from statistics is that it's useful to employ data when you make decisions. So, here are course evaluation statistics from my previous classes. Next, I especially enjoy teaching in a hybrid style. This requires some self discipline from students to watch or read lecture material outside of class, but the pay-off is that in class, we get to do the real learning: practice problems, trouble-shooting confusions with each other, and failing in a supportive environment (not just the students; I've learned so much from my failures as a teacher). It's also interesting to integrate into class ideas from audio/books on learning, such as Great Courses: The Learning Brain, by Professor Thad A. Polk, and A Mind for Numbers, by Barbara Oakley, Ph.D. Also, I am immensely grateful for various workshops that have been available to me through the CSU & CCC Math Learning Community, through the National Center for Faculty Development & Diversity, through SFSU's Center for Equity and Excellence in Teaching & Learning and USC's Center for Excellence in Teaching.
