Third Conference on Statistics and Data Science - CSDS 2021

Title: Moving to a World Beyond p<0.05
Abstract: For nearly a hundred years > the concept of “statistical significance” has been fundamental to statistics and to science. And for nearly that long, it has been controversial and misused as well. In a completely non-technical (and generally humorous) way, ASA Executive Director Ron Wasserstein will explain this controversy, and say why he and others have called for an end to the use of statistical significance as means of determining the worth of scientific results. He will talk about why this change is so hard for the scientific community to make, but why it is good for science and for statistics, and will point to alternate approaches.
Bio: Ronald L. (Ron) Wasserstein is the executive director of the American Statistical Association (ASA). Wasserstein assumed the ASA’s top staff leadership post in August 2007.
In this role, Wasserstein provides executive leadership and management for the association and is responsible for ensuring that the ASA fulfills its mission to promote the practice and profession of statistics. He also is responsible for a staff of 36 at the ASA’s headquarters in Alexandria, Va. As executive director, Wasserstein also is an official ASA spokesperson.
Prior to joining the ASA, Wasserstein was a mathematics and statistics department faculty member and administrator at Washburn University in Topeka, Kan., from 1984–2007. During his last seven years at the school, he served as the university’s vice president for academic affairs.
Wasserstein is a longtime member of the ASA, having joined the association in 1983, and has been active as a volunteer in the ASA for more than 20 years. He twice served as president of the Kansas-Western Missouri Chapter of the ASA. Wasserstein served as chair of two ASA sections—the ASA Section on Statistical Education and the ASA Section on Statistical Consulting. He also chaired the Council of Chapters Governing Board in 2006 and was a member of the ASA Board of Directors from 2001–2003.
Wasserstein is a Fellow of the ASA and American Association for the Advancement of Science. He was presented the John Ritchie Alumni Award and Muriel Clarke Student Life Award from Washburn University, the Manning Distinguished Service Award from the North American Association of Summer Schools, and the George Mach Distinguished Service Award from Kappa Mu Epsilon National Mathematics Honor Society.
Ron and his wife, Sherry, live in northern Virginia and enjoy jogging, movies, binge-watching TV series, live theater, books, and doting on their children and grandchildren.

Title: General Perception
Abstract: With ever growing datasets and computational power there seems to be ever diminishing value for embedding domain knowledge into the design of feature extractors. In this talk I will review some of the research done at DeepMind around video understanding and how that motivated our attempt at creating a scalable perception model that makes no architectural assumptions about its inputs: the Perceiver. I will discuss how domain knowledge can still be leveraged with the Perceiver and show results on various tasks across multiple modalities, including text, vision, audio and 3d geometry.
Bio: Joao Carreira has been a Research Scientist at DeepMind in London since 2016. Before that he received his Ph.D. from the University of Bonn in Germany in 2012 and was a postdoctoral fellow in UC Berkeley (2014-2015). Early on he worked on object segmentation and reconstruction, more recently in action recognition, video understanding and even more recently on general perception models -- highlights including the popular Kinetics dataset, I3D and Perceiver models.

Title: Family of mean-mixtures of multivariate normal distributions: Properties, inference and assessment of multivariate skewness.
Abstract: In this work, a new mixture family of multivariate normal distributions, formed by mixing multivariate normal distribution and a skewed distribution, is constructed. Some properties of this family, such as characteristic function, moment generating function, and the first four moments are derived. The distributions of affine transformations and canonical forms of the model are also derived. An EM-type algorithm is developed for the maximum likelihood estimation of model parameters. Some special cases of the family, using standard gamma and standard exponential mixture distributions, denoted by MMNG and MMNE, respectively, are considered. For the proposed family of distributions, different multivariate measures of skewness are computed. In order to examine the performance of the developed estimation method, some simulation studies are carried out to show that the maximum likelihood estimates do provide a good performance. For different choices of parameters of MMNE distribution, several multivariate measures of skewness are computed and compared. Because some measures of skewness are scalar and some are vectors, in order to evaluate them properly, a simulation study is carried out to determine the power of tests, based on sample versions of skewness measures as test statistics for testing the fit of the MMN E distribution. Finally, two real data sets are used to illustrate the usefulness of the proposed model and the associated inferential methods.
Bio: Narayanaswamy Balakrishnan is Distinguished University Professor of the McMaster University, Canada. He is Elected Member of the ISI, and fellow of the ASA and IMS. His current editorial activities include being Editor-in-Chief for Communications in Statistics-Theory and Methods, Communications in Statistics-Simulation and Computation, and Communications in Statistics-Case Studies, Data Analysis and Applications, and Associate Editor for Journal of Probability and Statistical Science, Methodology and Computing in Applied Probability, Metrika and Test. Bala has supervised more than 50 MSc thesis and more than 50 PhD thesis, published 20 books and hundreds of scientific papers, and has more than 81.000 citations on google scholar.

Title: DNA Copy Number Profiling from Bulk Tissues to Single Cells
Abstract: The completion of the human genome two decades ago gave birth to the expansive and cross-disciplinary field of Genomics, and along with it, our own community of Statistical Genomics. From microarrays to high throughput sequencing, from genome-wide association studies to the recent advances in single cell profiling, wave after wave of technological innovation have fed Statistics with new data challenges that spurred methodological and theoretical developments. In this lecture, I will focus on two specific topics in Genomics: single cell sequencing and DNA copy number profiling, and describe the critical role of Statistics in their scientific development. I will start with DNA copy number profiling in bulk tissues, review the scientific background and early models, and describe how these models have adapted to adjust to the shifting sands of technological change. I will briefly survey the statistical developments that were seeded by these scientific inquiries, from change-point detection to multi-channel scan statistics to latent variable modeling. On the scientific side, I will focus on DNA copy number profiling in cancer and its role in the study of cancer cell evolution.
Despite our best computational efforts, bulk tissue sequencing can only tell us so much about how DNA copy number varies between single cancer cells within a tumor. Cancer is a Darwinian evolution of cells driven by somatic mutations, and it is important to detect and study these cell-to-cell DNA copy number variations. In the second half of my talk, I will turn to the modeling of data from single cell technologies, which have revolutionized the field of biology during the last decade. I will describe how the large, sparse data matrices from single cell experiments have inspired new models and statistical problems. I will also describe, to some detail, a specific method that we developed for allele-specific copy number estimation at the single cell level. The method, Alleloscope [1], has enabled the discovery of previously hidden types of variation within tumor cell populations.
References: Wu C-Y, Lau BT, Kim H, Sathe A, Grimes SM, Ji HP, Zhang NR. Integrative single-cell analysis of allele-specific copy number alterations and chromatin accessibility in cancer. Nature Biotechnology, May 20, 2021. HTTPS://DOI.ORG/10.1038/S41587-021-00911-W .
Bio: Dr. Zhang is a Ge Li and Ning Zhao Professor of Statistics in The Wharton School at University of Pennsylvania. Her research focuses primarily on the development of statistical and computational approaches for the analysis of genetic, genomic, and transcriptomic data. In the field of Genomics, she has developed methods to improve the accuracy of copy number variant and structural variant detection, methods for improved FDR control in genomic studies, and methods for analysis of single-cell RNA sequencing data. In the field of Statistics, she has developed new models and methods for change-point analysis, variable selection, and model selection. Dr. Zhang has also made contributions in the area of tumor genomics, where she has developed analysis methods to improve our understanding of intra-tumor clonal heterogeneity.
Dr. Zhang obtained her Ph.D. in Statistics in 2005 from Stanford University. After one year of postdoctoral training at University of California, Berkeley, she returned to the Department of Statistics at Stanford University as Assistant Professor in 2006. She received the Sloan Fellowship in 2011, and formally moved to University of Pennsylvania in 2012. She was awarded the Medallion Lectureship by the Institute of Mathematical Statistics in 2021. At Penn, she is a member of the Graduate Group in Genomics and Computational Biology, and currently serves as the Vice Dean of the Wharton Doctoral Program.

Title: Reliability in Brazil: Roads for approaching industry
Abstract: Our dependence on mechanical and electronic devices is increasing. However, no matter how efficient they are, can fail. For instance, we can mention technologies embedded in intelligent sensors, artificial intelligence devices, agricultural, financial, and medical robots. In this context, statistical reliability analysis has been extensively used, inserted into innovation processes. At this conference, some reliability innovation projects are presented, showing how we are creating a connection road between academia and the industrial, medical and financial sectors. Focus is given on reliability modeling for oil well construction equipment, bucket tracking equipment, agricultural machinery, and communication modeling for mobile phones.
Bio: Francisco Louzada is a Professor of Statistics at the University of São Paulo (USP), and Director of the Center for Mathematics and Statistics Applied to Industry (CeMEAI). He obtained his Ph.D. in Statistics from the University of Oxford. His main interests are data science, risk analysis, and statistical inference and its applications. More details about Prof. Louzada can be found in www.grupocer.org/franciscolouzada.

Title: Statistical Leadership
Abstract: In this talk I want to set out what we mean by statistical leadership, and will consider the questions: who needs to demonstrate statistical leadership; why leadership is needed at all levels of a statistical organisation, and what makes a good statistical leader – why statisticians don’t always make good leaders; and how we can grow statistical leadership.
It is based on a lifetimes experience as an official statistician in the UK. But having talked to many statisticians on other countries, I feel it has a wider relevance. Nor are the issues particular to national statistics offices – they have relevance to wherever statisticians need to lead – in research institutes, universities, or businesses.
Bio: Stephen Penneck is President of the International Statistical Institute, and Honorary Officer for Official Statistics at the Royal Statistical Society, where he is a Fellow and a Chartered Statistician. Stephen retired in 2012 as Director General at the UK’s Office for National Statistics (ONS), following a career in official statistics.
Since retirement, Stephen has published articles and given lectures on a range of topics in official statistics including economic statistics, open data, and governance, trust and ethics; contributed to ISI workshops and regional conferences in Korea, Cameroun, Malaysia and Tunisia; and organised sessions at recent IAOS conferences and World Statistics Congresses.